Propositions · 24 propositions · 323 records · 1952–2026 Economics of machine intelligence Which located texts, from 1952 to 2026, state, contest or bear on propositions about the economics of machine intelligence? Three strands are recorded. The first is the debate over whether machine intelligence could make economic growth explosive. The second is how mainstream economics treated automation, computers and AI. The third is how economists took up recursive self-improvement, superintelligence and existential risk from AI, ideas mostly discussed outside economics. For each proposition, which text by an economist is the earliest located, and which text outside that group?
Each proposition has one criterion. It is phrased generically and attributed to no one. A text that meets the criterion articulates the proposition. A text that disputes it, or a premise it depends on, contests it, and the record says whether the passage gives a reason against a premise or states the contrary only. A text that comes close bears on it, and its record names what is missing. A result from a formal model is shown with the conditions under which the model derives it. Each text carries a tier: tier 1 for scholarly work by economists and for the core economics journals and working-paper series, tier 2 for other texts. Texts are listed in date order. The page records what texts say and when. It does not assess whether an argument or a model is right.
A machine selected and arranged these records. No human review of the selection or framing is recorded.
Download research starter The original sources first, in date order, then every proposition on this page with its criterion, earliest located text and records. Each record gives its original source and the place in it, its role, register and dates, the qualification and the missing element of a context record, then the Selekoa identifiers. Stored paraphrases and permitted short quotes are included. Original documents are not.
Download Markdown Download JSON
Original sources for a reference manager. Every source this export cites, for Zotero, EndNote and similar tools. Cite the original documents, not Selekoa, and check each entry against them.
Download BibTeX Download RIS Download CSL-JSON
Copy link to this page
G · The explosive-growth debate G01 Machine substitution for labour produces explosive growth G02 Automated research produces accelerating growth G03 Long-run history suggests a further acceleration G04 Essential unautomated tasks bottleneck growth G05 Other constraints prevent explosive growth G06 Current data show no approach to explosive growth G07 Machine substitution drives wages down G08 AI’s near-term aggregate productivity effect is modest G09 Estimates of explosive growth within a stated horizon M · Mainstream economics, automation and AI M01 Computers and AI as ordinary technical change M02 AI as the automation of tasks M03 AI as a general-purpose technology with lagging productivity M04 AI as cheaper prediction M05 Machines cheaper than labour at every task M06a Automation lowers the labour share, wages or employment M06b Compensating policy for AI’s distributional effects M07 Exposure to automation measured from task content M08 AI’s productivity effect over the next decade M09 Recent entry of machine intelligence into mainstream economics E · Superintelligence and existential risk E01 Machine intelligence improving itself accelerates its capability growth E02 Machine intelligence exceeding human cognition in essentially all domains E03 Existential risk from AI as a matter for economic analysis E04 AI developers’ incentives matter for safety E05 Economics’ engagement with superintelligence, AGI and the singularity How we searched The collection records what located texts from 1952 to 2026 say about the propositions below. “Earliest located” means the earliest text found by the searches below, within tier 1 unless the page says tier 2. It is not a claim that no earlier text exists.
Venues and years searched The explosive-growth debate, 1994 to 2024 Robin Hanson: “If Uploads Come First” (1994), “Is a Singularity Just Around the Corner?” (1998, read for this strand on 2026-10-25), his comment on Vernor Vinge’s singularity concept (1998), “Economic Growth Given Machine Intelligence” (undated, attested 2000) and “Long-Term Growth As A Sequence of Exponential Modes” (2000). William D. Nordhaus, “Are We Approaching an Economic Singularity?” (2015); Philippe Aghion, Benjamin F. Jones and Charles I. Jones, “Artificial Intelligence and Economic Growth” (2017); Philip Trammell and Anton Korinek, “Economic Growth under Transformative AI” (2020, text of 2021); Daron Acemoglu, “The Simple Macroeconomics of AI” (2024). David Roodman, “Modeling the Human Trajectory” (2020), with Holden Karnofsky’s comment on it; Tom Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021); the Asterisk debate between Matt Clancy and Tamay Besiroglu (2023); Ege Erdil and Tamay Besiroglu’s review of the arguments (arXiv, 2023). Texts were chosen from a starting list and from the texts these cite or answer directly. Each was read in an Internet Archive capture or a dated arXiv version, retrieved on 2026-10-08 and later re-checked. Mainstream economics, 1952 to 2024 A census of titles: every Crossref record registered under the AER, QJE, JPE, Econometrica, REStud, JEP, JEL and AEA Papers and Proceedings, and under the NBER DOI prefix, 81,884 records, matched against terms for AI, machine learning, automation and computers. Crossref coverage is uneven: AER and JEL records start in 1999 and JEP in 1987. Full texts of NBER working papers on AI and automation, four chapters of the NBER volume The Economics of Artificial Intelligence (2019), Frey and Osborne (2013), Eloundou and others (arXiv, 2023), Nilsson (AI Magazine, 1984) and Bresnahan, Brynjolfsson and Hitt (1999). Agrawal, Gans and Goldfarb’s essay in the Harvard Business Review (2016) was added on 2026-10-25. Abstract pages only for articles in the JEP, the AER and AEA Papers and Proceedings, whose full texts were refused, and the publisher’s standfirst only for Leontief’s “Machines and Man” (Scientific American, 1952). The 1955 congressional hearings on automation and technological change, the 1966 report of the National Commission on Technology, Automation, and Economic Progress with its second appendix volume, Bowen and Mangum’s 1966 condensation, and Leontief and Duchin’s 1984 report to the National Science Foundation, read in Internet Archive text layers. Texts were retrieved directly from publishers, NBER, arXiv and the Internet Archive on 2026-10-09, as current copies. An archive capture was not available, so a current copy does not authenticate earlier bytes. The Finnish strand Kansantaloudellinen aikakauskirja: the archive of 147 issues was crawled on 2026-10-09 and searched in Finnish and English terms for automation, computers and AI (2,854 PDFs; 32 without a text layer). Issues from 3/2023 to 3/2026 were searched through their titles and abstracts only. The crawl found 247 texts on automation, computers or AI, listed with checked quotations in the research pack (research/economics-mainstream/kak.md). They were classified by subject, not against the propositions here, so they are not records on this page. ETLA Brief 22 (Pajarinen and Rouvinen, 2014), read in its abstract, is recorded under M07. Superintelligence and existential risk, 1998 to 2026 Robin Hanson: “Is a Singularity Just Around the Corner?” (Journal of Transhumanism, 1998), his comment on Vernor Vinge’s singularity concept (1998) and “Economic Growth Given Machine Intelligence” (undated, attested 2000), read in Internet Archive captures. Posts from his 2008 debate with Eliezer Yudkowsky and two later posts, read in the Machine Intelligence Research Institute’s compilation of 2013; the date of “AI Go Foom” was checked on the original page. The NBER working papers on AGI, transformative AI or existential risk from AI that the title census found (2018 to 2026), among them Gans (2018, 2023), Jones (2023, 2025), Korinek and Suh (2024), and the October 2023 version and April 2026 revision of Trammell and Korinek, whose wording differs. Five papers of 2025 and 2026, whose full texts were refused (HTTP 403), and Jones’s article in AER: Insights (2024) were read as abstract pages only. Texts already in the collection, re-read for these propositions: Sachs and Kotlikoff (2012), Nordhaus (2015), Aghion, Jones and Jones (2017), Korinek and Stiglitz (2017), the introduction and Tyler Cowen’s chapter of the NBER volume (2019) and Acemoglu (2024). Ari Hyytinen’s editorial in Kansantaloudellinen aikakauskirja (4/2017). The title census of the mainstream strand, extended to AER: Insights, was searched for terms such as superintelligence, singularity, intelligence explosion, AGI, existential risk and takeoff, and the Finnish texts for their Finnish equivalents. Texts were retrieved on 2026-10-09 and 2026-10-11 and, except the Internet Archive captures, retrieved again on 2026-10-15 and 2026-10-16. Methods Machine searches read each text proposition by proposition: abstracts, introductions, conclusions and objection sections in full, and long texts by targeted search. Every quotation was compared by script with the checked copy of its text, after normalising spacing, quotation marks, dashes and line-break hyphenation. Quotations are at most 50 words, and the quoted words from one document stay within 400, or about a tenth of a short text. Each text is dated by its own date line, cover or title page. Where the quoted wording is known only from a later copy, such as an archive capture or a revision of a working paper, the record also gives the date from which that wording is attested, and that date decides which text is earliest located. A journal year never dates a working paper’s wording. Tiers follow the maintainer’s rule of 2026-10-11, with an author’s standing as an economist judged when the text was written (2026-10-12). Each record gives the basis of its tier. A passage engages a premise when its section, dialogue turn or message gives a reason against that premise. Otherwise it is recorded as stating the contrary. Results from formal models are recorded with their conditions. Model scenarios counted as articulating M05 are labelled as such and shown only with their context (maintainer decision of 2026-10-11). Estimates under G09 are recorded one per text and holder, as worded, and are never converted to a common threshold. Fresh-context audits by Claude models re-read every passage. Checks by an OpenAI GPT-6 model then re-examined 49 of 65 passages of the explosive-growth strand and 62 passages and four Finnish texts of the mainstream strand. Correction passes applied their findings on 2026-10-10 and 2026-10-12; for the mainstream strand the maintainer accepted every final disagreement. No person has reviewed the records. The records were admitted on 2026-10-14 from two research packs. One quotation was shortened to keep it within 50 words; its wording is unchanged. On 2026-10-25 a same-family fresh-context machine check re-examined every tier and every earliest located role of the first two strands, and a blind cross-family machine check re-examined 19 of its items against the recorded captures and the hearing scans. The maintainer decided their disagreements the same day: one text was re-tiered, six records were added, five earliest located roles moved and one was withheld. In that correction pass web.archive.org refused connections, so passages read only in archive captures rest on the GPT check. For the strand on superintelligence and existential risk, an OpenAI GPT-6 model checked all 119 passages blind on 2026-10-15 and found every quotation matching. Its 22 disagreements on classification were decided on the evidence the same day, and the maintainer decided on 2026-10-15 how E05 is read. The 20 passages read in Internet Archive captures could not be read again at admission, because the archive refused connections; they rest on that check, which retrieved the same captures. The strand was admitted on 2026-10-16. Three passages were held back to keep within the quotation limit for Hanson’s 1998 comment, and one quotation was shortened to 50 words with its wording unchanged. Known gaps I. J. Good’s 1965 chapter could not be read in either archive capture, which are truncated. Good appears only as later texts relay him. Herbert Simon’s The Shape of Automation (1965) was not retrieved. Nordhaus (2015) and Armer (1966) relay its forecast that computers could do any work a person can do, and Simon may state M05 in his own voice. The journal versions of the working papers were not compared, among them Acemoglu’s article in Economic Policy (2025), whose publisher refused requests (HTTP 403). Whether that version changes the M08 figures is unknown. A review of 1931 in the Journal of Political Economy whose title names robots could not be retrieved (HTTP 403), so what “robots” means in it is unknown. Roodman’s technical paper behind his 2020 post was not retrieved. Davidson’s report and its appendices were searched, not read in full. Hanson’s 1994 essay was read in a web reprint captured in 1998. The printed issue of Extropy, which the dataset holds as a scan, was not compared with the quoted wording. Full texts of JEP, AER and AEA Papers and Proceedings articles, of JSTOR and of Google Books were not reachable. The AER and the JEL before 1999 were not covered at title level. OpenAlex, the IMF and the OECD were not searched. Books, podcasts, social media and comment threads were not searched, apart from one comment recorded under G03 and G09. The non-English literature was searched only in Kansantaloudellinen aikakauskirja and one ETLA search. A text found later that is earlier than an earliest located text, or written by an economist, would change the headlines. Twenty-one texts not retrieved are listed as leads in the research pack, among them Simon (1965) and Kai Husso’s 1992 chapter on the diffusion of AI in Finland. G09 has no earliest located tier-1 text for now (maintainer decision of 2026-10-25). Hanson’s “Long-Term Growth As A Sequence of Exponential Modes” (December 2000; first version September 1998), already recorded under G09, is an unchecked lead for the role: its attested wording and its link to machine intelligence have not been compared for this purpose, and the 1998 first version was not retrieved. For M05, Hanson’s “Economic Growth Given Machine Intelligence” models machines doing most jobs but says: “We have not, however, considered the possibility of creating new kinds of jobs” (p. 12), so it does not meet M05’s condition on tasks created later. Leontief’s “The Distribution of Work and Income” (Scientific American, 1982), which that paper cites, is an open lead; its text could not be retrieved. Institutional forecasts, such as those of Goldman Sachs (2023) and the McKinsey Global Institute (2023), both cited by Acemoglu (2024), give earlier ten-year estimates that would bear on M08. They are outside the collection’s scope until a rule for selecting institutional sources exists (maintainer decision of 2026-10-25). G06 has no tier-1 articulating text under its literal wording. Nordhaus (2015) reports two positive tests of seven and concludes that the singularity is not near, so his record bears on G06. Rewording the criterion is an open decision. M09’s tier-1 articulating record is a self-description of 2019. It must be read with the dated record from 1952 onward, which runs against a plain reading that economics took up the subject only recently. E05’s earliest located tier-1 text is Nordhaus (2015), a statement about the state of the field. It must be read with the records under E01 to E04, which show economists engaging with these ideas from 1998. For the strand on superintelligence and existential risk, books were not searched, among them Hanson’s The Age of Em (2016) and Cowen’s Average Is Over (2013), nor Hanson’s article in IEEE Spectrum (2008), economics blogs beyond the compiled debate, or journals outside the title census. Puhakka’s editorial in Kansantaloudellinen aikakauskirja 3/2024 was not read. A text found later that is earlier than the earliest located text becomes the new earliest located text. The earlier headline stays on record as a correction.
How to read a record Articulates A passage meets the proposition’s criterion. Contests A passage disputes the proposition or a premise it depends on. Bears on A passage comes close without meeting the criterion; the record names what is missing. Earliest located The earliest articulation the searches found, dated by its wording. It is not a claim that no earlier text exists. The earliest located text that contests a proposition is chosen by the same rule. Tier 1, tier 2 Tier 1: Scholarly work by economists in any venue, including their essays in magazines, and any text in a core economics venue, whoever wrote it. An economist has an economics appointment, degree, address or affiliation shown in or near the text, or an established record as an economist when the text was written. The core venues are the five leading general economics journals, the Journal of Economic Perspectives, the Journal of Economic Literature, AER Papers and Proceedings, NBER working papers and the NBER Economics of AI volume, and Kansantaloudellinen aikakauskirja for the Finnish texts. Tier 2: Texts by non-economists, whatever the venue, and short-form commentary even by economists: blog comments, forum replies, and magazine or online debates, unless a turn is itself a substantive scholarly argument. Earliest located is named within tier 1, and the earliest located tier-2 text separately. A tier is not a ranking of quality. Earliest located use of the name The earliest located text that uses the proposition’s usual name. A title can carry the name without stating the proposition. Conditions The model assumptions or stated conditions under which a passage holds. A conditional result is not a forecast. Model scenario A case of a formal model, under its conditions, that meets the criterion. It is not the author’s own claim, and it is shown only with its context. Estimate An estimate as the text states it, with its holder, horizon and the text’s own definition. Estimates from different texts are not converted or combined. Engages a premise, states the contrary A contesting passage engages a premise when its paragraph, message or dialogue turn gives a reason against it. Otherwise it states the contrary conclusion only. Shown only with its context A passage that would mislead on its own, such as an earlier view without the later text that rejects it. The evidence inspector shows it only together with the context it needs. Dates The first date is the one the work is dated by. “Wording attested” gives the date from which the quoted wording is known, when it survives only in a later copy. “Claimed” gives an earlier talk, draft or presentation that does not date the wording. Register: the kind of text Fiction, satire, a dream frame or a relay of another author is labelled. A passage in fiction records what the work says, not its author’s view. Unlabelled records are nonfiction. Limits of this page “Earliest located” means the earliest text found by the searches described under “How we searched”, by economists unless the page says tier 2. It is not a claim that no earlier text exists. A record documents what a text says. It is not evidence that the proposition is true or that a model is right, and no record links one text to another. A tier records who wrote a text and in which form, judged when it was written. It is not a ranking of quality. A model result holds only under its recorded conditions. An estimate is its author’s, under the author’s definition and horizon, and estimates from different texts are not comparable without them. Each record carries the date its work is dated by. Where the quoted wording is known only from a later copy or revision, the date from which it is attested is shown, and that date decides which text is earliest located. A passage that would mislead on its own is marked “shown only with its context” and is never shown without that context. Codes such as M05-20 in qualifications refer to the passages of the research packs the records come from. A machine selected and classified these records, and machine checks from two model families re-examined them. No person has reviewed them. G · The explosive-growth debate Whether machine intelligence could make growth far faster than historical rates, the constraints argued against it, the evidence offered on both sides, and each text’s own estimate. Thirteen texts from 1994 to 2024.
G01 Machine substitution for labour produces explosive growthCriterion. The text states that if machines (AI systems or emulated minds) can substitute for human labour across all or nearly all tasks, so that labour can be produced like capital, the growth rate of output rises far above historical rates (by roughly an order of magnitude or more) or keeps rising. The conditions under which the text derives this are part of the record.
Earliest located, tier 1 Robin Hanson · Economic Growth Given Machine Intelligence · 2000-08-19 · wording attested 2000-08-19 · claimed 1998 ↗
Earliest located, tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? · 2021-06-25 · wording attested 2021-07-02 ↗
Earliest located tier-2 text that contests it Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth · 2023-06 · wording attested 2023-12-23 · claimed 2023 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (7) Articulates · Earliest located · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) articulates G01. Summarises the paper's exogenous-growth models: once machines substitute for human labour, the labour force can be produced like capital and growth rates rise. The paper's numerical example gives 4.3% growth before and 45% after. States the two assumptions the result depends on. Conditions: Exogenous technology growth; computer prices halving every two years; machine labour instantly producible; diminishing returns to all inputs together; product shares chosen as "conservative" (labour 0.25, capital 0.5, computers 0.02). 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates G01. States the model condition for a supply-side "Singularity": substitution of information capital for labour above unit elasticity. The paper then tests for the signs this mechanism would leave. Conditions: Two-input production with information capital and labour; constant elasticity of substitution above one; information capital improving rapidly. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates G01. With every task in goods production automated, the model becomes an AK model and ongoing productivity growth makes growth rates themselves rise over time ("Type I": growth rates rising without bound). Conditions: Full automation of goods production; ongoing technological progress. 2017-10 · wording attested 2018-12-14 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 · Model scenario Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) articulates G01. In a toy scenario in which AI robots substitute perfectly for all human labour, labour can be accumulated by investing in robots, and the report concludes that growth models then make super-exponential growth the default expectation. Conditions: A toy scenario in which AI robots can perfectly substitute for all human labour, so that labour can be accumulated through investment in robots. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G01. Speaker: Tamay Besiroglu. Besiroglu defines explosive growth as at least ten times the past century's annual growth rate, sustained for at least a decade, and gives a conditional probability. Conditions: Conditional on AI automating the vast majority of tasks; explosive growth defined as at least tenfold the past century's growth rate for a decade or more. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) contests G01. Speaker: Matt Clancy. Clancy questions whether growth models can be extrapolated to growth far outside historical experience. Earlier in the debate he puts the chance of explosive growth at 10–20%. Clancy's answer to Besiroglu's 65%, given on the same condition of AI deployed throughout the economy. He adds that most of the probability is "clustered around the low end of explosive growth", and the same turn doubts that growth models can be extrapolated so far. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ege Erdil; Tamay Besiroglu · Explosive growth from AI automation: A review of the arguments Erdil and Besiroglu, “Explosive growth from AI automation: A review of the arguments” (2023-10-01) articulates G01. Opens the paper's summary of its assessment: substitution of AI for labour on most or all tasks makes explosive growth the default prediction of growth models. Conditions: AI able to substitute for human labour on most or all tasks. 2023-10-01 · wording attested 2023-10-01 · claimed 2023-09-20 Inspect claim & evidence ↗ Later texts by the same author that contest it Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth · 2023-06 · wording attested 2023-12-23 · claimed 2023 ↗ Document page → Texts that bear on it without meeting the criterion (2) Bears on · Tier 2 Robin Hanson · If Uploads Come First: The Crack of a Future Dawn Hanson, “If Uploads Come First: The crack of a future dawn” (1994-03-08) bears on G01. If copying uploads is fast and cheap, their population could outgrow total wealth, lowering per-copy wealth. Missing: States that the population of uploads grows faster than wealth; does not state that the growth rate of total output rises. 1994-03-08 · wording attested 1998-02-20 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") Hanson, “Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept")” (1998-10) bears on G01. In a comment that is otherwise sceptical of Vinge's singularity, Hanson accepts that uploads could bring bursts of growth by removing biological limits on reproducing "human" capital. Missing: Describes temporary spurts of growth, not a sustained rise in the growth rate by an order of magnitude. 1998-10 · wording attested 2000-11-17 Inspect claim & evidence ↗ Document page → Headlines replaced Erdil and Besiroglu’s review (2023), headlined as the earliest located tier-2 text. Re-tiered as tier 1 after the re-checks of 2026-10-25: Besiroglu held an economics degree before the text. Davidson’s report (2021), whose toy scenario of AI robots substituting for all human labour meets the criterion, is now the tier-2 earliest. Recorded 2026-10-25. G02 Automated research produces accelerating growthCriterion. The text states that if machines substitute for, or improve, the producers of ideas (researchers, inventors, engineers), feedback from output to research effort makes growth accelerate: super-exponential growth, growth rates rising without bound, or a finite-time singularity. The conditions under which the text derives this are part of the record.
Earliest located, tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth · 2017-10 · wording attested 2018-12-14 ↗
Earliest located, tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? · 2021-06-25 · wording attested 2021-07-02 ↗
Earliest located tier-1 text that contests it Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth · 2017-10 · wording attested 2018-12-14 ↗
Earliest located tier-2 text that contests it Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") · 1998-10 · wording attested 2000-11-17 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (5) Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") Hanson, “Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept")” (1998-10) contests G02. Answers the claim that smarter entities produce even smarter entities faster: improvement gets harder as easy gains are used up, as with schooling, factory learning and the Flynn effect. Argues that greater intelligence or speed changes the level of world product once, not its later growth rate. The next sentences grant that growth may rise with the size of the economy, as it has for millennia. 1998-10 · wording attested 2000-11-17 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates G02. If an AI replaces all people in producing ideas and ideas build on earlier ideas, the model reaches infinite income in finite time. Conditions: All tasks in idea production automated; returns to existing knowledge in idea production (φ) above zero. 2017-10 · wording attested 2018-12-14 Inspect claim & evidence ↗ Document page → Contests · Earliest located · Tier 1 · Engages a premise Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) contests G02. If ideas become harder to find fast enough, even fully automated research does not produce a singularity. 2017-10 · wording attested 2018-12-14 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) articulates G02. Restates the ideas feedback loop of semi-endogenous growth models with AI systems in place of population, after the demographic transition broke the human loop. Draws the conclusion of the feedback argument (audit A: the first passage alone stops before it). Conditions: AI substituting very effectively for human labour in both production and R&D. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G02. Speaker: Tamay Besiroglu. Besiroglu argues that AI restores the population-ideas feedback that drove historical acceleration. Conditions: AI that substitutes for human workers and idea producers; a semi-endogenous growth model in which ideas are non-rival and output, population and ideas feed back on each other. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (2) Bears on · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) bears on G02. In a variant where technology improves with production experience, some parameter values remove diminishing returns and growth keeps accelerating. The author sets this aside to stay "conservative". Missing: The feedback runs through learning by doing in production, not through machines doing research. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI: A guide to the vast range of possibilities for output growth, wages, and the labor share Trammell and Korinek, “Economic Growth under Transformative AI: A guide to the vast range of possibilities for output growth, wages, and the labor share” (2021-02-04) bears on G02. Infers from historical data that returns to existing knowledge in research have been below the threshold at which a constant research effort produces a finite-time singularity. Missing: Concerns the historical value of a parameter; it does not say whether AI researchers, whose number can grow, would produce accelerating growth. 2021-02-04 · wording attested 2021-06-25 · claimed 2020-10 Inspect claim & evidence ↗ Document page → Headlines replaced Hanson’s comment on Vinge (1998), headlined as the earliest located text that contests it before texts were sorted into tiers. Short-form commentary is tier 2 under the rule of 2026-10-11, so it is now the tier-2 earliest. Aghion, Jones and Jones (2017) is the earliest located tier-1 text that contests it. Recorded 2026-10-11. G03 Long-run history suggests a further accelerationCriterion. The text infers from the long-run history of world output or population (hyperbolic or super-exponential growth, or a sequence of ever faster growth modes) that a future transition to much faster growth is plausible or expected.
Earliest located, tier 1 Robin Hanson · Is a Singularity Just Around the Corner? What it takes to get explosive economic growth · 1998-06 · wording attested 1999-01-28 · claimed 1998-03-30 ↗
Earliest located, tier 2 Holden Karnofsky · Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece") · 2020-06-15 · wording attested 2020-06-23 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (7) Articulates · Earliest located · Tier 1 Robin Hanson · Is a Singularity Just Around the Corner? What it takes to get explosive economic growth Hanson, “Is a Singularity Just Around the Corner? What it takes to get explosive economic growth” (1998-06) articulates G03. From historical trends in growth rates he infers that the economy is slowly moving towards a high-growth mode and may reach it in roughly a century and a half. Conditions: The Conclusion opens: "If we assume that the risk and time preferences of future investors will not vary much from those which evolution has given us". 1998-06 · wording attested 1999-01-28 · claimed 1998-03-30 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) articulates G03. Cites the author's own analysis of long-run growth (the 1998 first version of "Long-Term Growth As A Sequence of Exponential Modes") against rejecting the model's fast growth "out of hand". 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Robin Hanson · Long-Term Growth As A Sequence of Exponential Modes Hanson, “Long-Term Growth As A Sequence of Exponential Modes” (2000-12) articulates G03. Fits world product over two million years with a few exponential "modes" (hunting, farming, industry) and extrapolates the pattern of transitions. Conditions: The sum-of-exponentials or CES-combined model of growth modes; parameters of a new transition similar to those of past transitions. 2000-12 · wording attested 2001-11-28 · claimed 1998-09 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Holden Karnofsky · Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece") Karnofsky, “Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece")” (2020-06-15) articulates G03. Karnofsky, who writes that he asked Roodman to evaluate Hanson's work, states what he takes from the post; he "wouldn't take the model discussed in this piece literally". 2020-06-15 · wording attested 2020-06-23 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 David Roodman · Modeling the Human Trajectory Roodman, “Modeling the Human Trajectory” (2020-06-15) articulates G03. Roodman fits a stochastic super-exponential model to gross world product since prehistory and draws out its instability. 2020-06-15 · wording attested 2020-06-23 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) articulates G03. An argument from history without a specific model: growth has sped up before, so a further increase should not be ruled out. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G03. Speaker: Tamay Besiroglu. Points to faster growth after the Neolithic and Industrial Revolutions, and to double-digit catch-up growth, as precedent. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (2) Bears on · Tier 2 Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") Hanson, “Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept")” (1998-10) bears on G03. Grants that growth rates may rise with the size of the world economy, but treats that as the long-standing accelerating trend, not a reason to expect a sudden break. Missing: Uses the accelerating historical trend to bound expectations; does not infer a transition to much faster growth. 1998-10 · wording attested 2000-11-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) bears on G03. Speaker: Matt Clancy. Clancy questions whether past accelerations are informative about an acceleration driven by intelligence. Missing: Disputes that past accelerations are informative about an acceleration driven by intelligence; G03 does not require that mechanism, and the same turn grants that past acceleration "is a good reason to think there is some chance it could happen again". 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Headlines replaced Hanson’s undated “Economic Growth Given Machine Intelligence” (attested 2000-08-19), headlined as earliest located from 2026-10-10. Earlier text found by the re-checks of 2026-10-25: Hanson’s “Is a Singularity Just Around the Corner?” (June 1998, attested 1999-01-28) makes the inference itself; the 2000 paper relays his own 1998 projection. Recorded 2026-10-25. Hanson’s “Long-Term Growth As A Sequence of Exponential Modes” (2000), headlined as earliest located in the first pass. Earlier text found by the audit: Hanson’s undated “Economic Growth Given Machine Intelligence”, attested 2000-08-19. Recorded 2026-10-10. The Asterisk debate (2023), headlined as the earliest located text that contests it in the first pass. Reclassified as bearing on it after the cross-family check: the same turn grants that past acceleration is a reason to think it could happen again. Recorded 2026-10-10. G04 Essential unautomated tasks bottleneck growthCriterion. The text states that growth is constrained because essential tasks or inputs remain unautomated or unimproved, or are automated only slowly, so that automating many but not all tasks, or automating them gradually, does not by itself produce explosive growth (Baumol's cost disease; an elasticity of substitution below one between tasks or between information and other inputs). Reasons why automation stays incomplete or slow (technical difficulty, physical-world tasks, training data, regulation of AI training or deployment, slow adoption) belong here.
Earliest located, tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth · 2015-09 · wording attested 2016-08-09 ↗
Earliest located, tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth · 2023-06 · wording attested 2023-12-23 · claimed 2023 ↗
Earliest located tier-1 text that contests it Robin Hanson · Economic Growth Given Machine Intelligence · 2000-08-19 · wording attested 2000-08-19 · claimed 1998 ↗
Earliest located tier-2 text that contests it Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? · 2021-06-25 · wording attested 2021-07-02 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (7) Contests · Earliest located · Tier 1 · States the contrary Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) contests G04. In the paper's example, a residue of human-only jobs lengthens the economic doubling time from 18 to 27 months, against 16 years without machine intelligence: growth stays far above historical rates. Conditions: Product shares fixed; human-only jobs keep a share of 0.05 of output in the example. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates G04. Rapid improvement in information technology raises aggregate growth only if information substitutes well for other inputs in production, or for other goods in consumption. The inelastic case of the same model: growth converges to that of the slowest input. Conditions: (G04-a) Elasticity of substitution between information and other inputs or goods below one. (G04-b) Elasticity of substitution between information capital and labour below one. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates G04. The paper's organising theme, developed in its task-based models of automation. States the condition directly. The same example then shows that with Cobb-Douglas production a singularity can occur without full automation. Conditions: (G04-c) Tasks are complements (elasticity of substitution below one). (G04-d) CES production across tasks with elasticity of substitution below one. 2017-10 · wording attested 2018-12-14 Inspect claim & evidence ↗ Document page → Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) contests G04. Responds to the Aghion, Jones and Jones bottleneck. Davidson accepts that it "correctly highlights that AI may lead to very widespread automation without explosive growth occurring". 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu (2023-06) articulates G04. Speaker: Matt Clancy. Clancy gives a verbal version of the Aghion, Jones and Jones task model applied to innovation: with every task required and automation advancing about 1% a year, growth stays steady; only faster or full automation gives explosive growth. Clancy says AI is "pretty likely to boost growth" and then lists bottlenecks that slow AGI's takeover of tasks: physical-world tasks, raw materials, training data, the time natural processes take, preferences for human performers, regulation and zero-sum competition. Clancy's closing turn: different sectors are held back by materials, time, data or regulation; sectors that automate easily become cheap and "no longer count for much of GDP". Conditions: Every innovation task is required; innovation gets harder each year; displaced workers retrain onto the remaining tasks. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Contests · Tier 2 · Engages a premise Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) contests G04. Speaker: Tamay Besiroglu. Besiroglu argues that with complementary tasks, scaling digital workers on automated tasks still raises output by one or two orders of magnitude unless a large share of tasks stays unautomated. Conditions: Tasks are complements; "digital workers" on automated tasks scaled up at least tenfold; automation going "from little to substantial" over decades rather than about a century; substitution of AI for human labour not unusually hard. In the same turn Besiroglu says that with 25% of tasks unautomated output could still rise by one or two orders of magnitude. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · Engages a premise Ege Erdil; Tamay Besiroglu · Explosive growth from AI automation: A review of the arguments Erdil and Besiroglu, “Explosive growth from AI automation: A review of the arguments” (2023-10-01) contests G04. Based on a CES task model calibrated with a range of substitution parameters: fast automation of most tasks can give a large transitory boost even with bottlenecks. Section 3.1 treats the objection that "the training or deployment of AI systems will be sufficiently impeded by regulation"; the authors argue that incentives for AI adoption resemble those for nuclear proliferation, so coordination to slow it is likely to be imperfect. Conditions: The sentence opens "even if the other arguments fail to go through and we cannot even attain AI that perfectly substitutes for humans across all tasks". CES task model with complementary tasks; a fixed stock of inputs reallocated from automated to unautomated tasks; input to automated tasks approximated as infinite (very cheap, very large AI inputs), so finite AI input would reduce the effect; about 90% of tasks automated (f = 0.1 unautomated) within less than ten years; substitution parameter ρ = −2 between tasks, below the range of −1.2 to −0.14 the authors cite as plausible for capital and labour. 2023-10-01 · wording attested 2023-10-01 · claimed 2023-09-20 Inspect claim & evidence ↗ Document page → G05 Other constraints prevent explosive growthCriterion. The text states that constraints that would still bind if machines performed all or nearly all tasks (physical laws or limits, natural resources and energy, the time that natural processes and experiments take, regulation of activities other than AI training and deployment, or demand for human-provided goods) can prevent or quickly end explosive growth.
Earliest located, tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth · 2017-10 · wording attested 2018-12-14 ↗
Earliest located tier-1 text that contests it William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth · 2015-09 · wording attested 2016-08-09 ↗
Earliest located tier-2 text that contests it Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? · 2021-06-25 · wording attested 2021-07-02 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (5) Contests · Earliest located · Tier 1 · Engages a premise William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) contests G05. Raises the objection that energy, water, minerals and thermodynamics limit accelerating growth, and answers it with the fall in energy needed per computation since ENIAC. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates G05. Gives physical limits, such as Carnot's theorem for energy conversion, as examples of essential inputs whose productivity cannot keep rising. Conditions: Some essential input subject to physical limits on productivity; tasks are complements. 2017-10 · wording attested 2018-12-14 Inspect claim & evidence ↗ Document page → Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) contests G05. Responds to a list of bottlenecks (regulation, materials, physical experiments, human adjustment, fundamental limits) with incentives to remove them, past 10% catch-up growth and 30% growth of individual businesses. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Contests · Tier 2 · Engages a premise Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) contests G05. Speaker: Tamay Besiroglu. Besiroglu sets three conditions a resource bottleneck must meet (vital, hard to substitute, impossible to expand by an order of magnitude); he also argues that experiments can be parallelised and that regulation faces strong incentives to be bypassed. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · Engages a premise Ege Erdil; Tamay Besiroglu · Explosive growth from AI automation: A review of the arguments Erdil and Besiroglu, “Explosive growth from AI automation: A review of the arguments” (2023-10-01) contests G05. Accepts eventual limits and contests their timing, with an estimate of returns to scale. 2023-10-01 · wording attested 2023-10-01 · claimed 2023-09-20 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (1) Bears on · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI: A guide to the vast range of possibilities for output growth, wages, and the labor share Trammell and Korinek, “Economic Growth under Transformative AI: A guide to the vast range of possibilities for output growth, wages, and the labor share” (2021-02-04) bears on G05. Treats growth-rate increases and singularities as benchmarks that hold "for some time" before limits such as land or energy bind. Missing: Says limits end super-exponential growth eventually; does not say they prevent explosive growth or end it quickly. 2021-02-04 · wording attested 2021-06-25 · claimed 2020-10 Inspect claim & evidence ↗ Document page → Headlines replaced Davidson’s report (2021), headlined as the earliest located text that contests it in the first pass. Earlier text found by the audit: Nordhaus (2015) answers the physical-limits objection. Davidson’s report is tier 2 and is now the tier-2 earliest. Recorded 2026-10-10. G06 Current data show no approach to explosive growthCriterion. The text states, on the basis of current or recent economic data, that the economy shows no sign yet of approaching a growth singularity or explosive growth.
Earliest located, tier 1 No located tier-1 text meets the criterion.
Earliest located, tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? · 2021-06-25 · wording attested 2021-07-02 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (2) Articulates · Earliest located in tier 2 · Tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) articulates G06. Responding to the objection that macroeconomic data show none of the signs explosive growth would leave (the objection cites Nordhaus (2021)), Davidson accepts the absence of signs for the next two decades and limits its reach beyond. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G06. Speaker: Matt Clancy. Clancy answers Besiroglu's point about fast growth in AI training compute. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (1) Bears on · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on G06. Six diagnostic tests on US data (productivity growth, capital share, capital prices, capital-output ratio, information capital share, wage growth) plus a demand-side test. For the two positive tests, simple extrapolation puts the crossing point 100 years or more away. Missing: (G06-a) Nordhaus reports two of seven tests positive for Singularity; he concludes that it is "not near", not that the data show no sign of approach. Under the literal criterion ("no sign yet") the record bears on G06. (G06-b) Nordhaus reports two of seven tests positive for Singularity; he concludes that it is "not near", not that the data show no sign of approach. Under the literal criterion ("no sign yet") the record bears on G06. Conditions: Singularity defined in the paper as "a time when the economic growth rate crosses 20% per year"; tests on US data to about 2014. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Headlines replaced Nordhaus (2015), headlined as earliest located in the first pass. Reclassified as bearing on it after the cross-family check: he reports two positive tests of seven and concludes that the singularity is not near, which is not “no sign yet”. No tier-1 text now articulates G06. Recorded 2026-10-10. G07 Machine substitution drives wages downCriterion. The text states that once machines substitute for human labour at falling cost, the market wage of human labour falls (towards the cost of machine labour, or below subsistence), even while total output or wealth rises.
Earliest located, tier 1 Robin Hanson · Economic Growth Given Machine Intelligence · 2000-08-19 · wording attested 2000-08-19 · claimed 1998 ↗
Earliest located, tier 2 Robin Hanson · If Uploads Come First: The Crack of a Future Dawn · 1994-03-08 · wording attested 1998-02-20 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (3) Articulates · Earliest located in tier 2 · Tier 2 Robin Hanson · If Uploads Come First: The Crack of a Future Dawn Hanson, “If Uploads Come First: The crack of a future dawn” (1994-03-08) articulates G07. If brain emulations can be copied cheaply and selection favours those willing to work for little, copying continues until wages approach subsistence. Names human wages explicitly, for the period in which the upload population expands. Conditions: Uploads can be copied fast and cheaply; values evolve under selection; copying is not forcibly prevented. 1994-03-08 · wording attested 1998-02-20 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) articulates G07. In the paper's model, wages first rise and later fall at the rate of computer price declines, while growth rates rise. States the extreme case of the model's wage path. Conditions: Exogenous-growth model with machine labour substituting for human labour; falling computer prices. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates G07. A polar case of his model with heterogeneous labour; skilled labour, an imperfect substitute, fares differently. Conditions: Unskilled labour a perfect substitute for information capital; skilled labour highly but imperfectly substitutable. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Headlines replaced Hanson’s “If Uploads Come First” (1994), headlined as the earliest located tier-1 text on 2026-10-11. An author’s standing is judged when the text was written (2026-10-12). In 1994 he held no economics degree or appointment, so the essay is tier 2 and is now the tier-2 earliest. Recorded 2026-10-12. G08 AI’s near-term aggregate productivity effect is modestCriterion. The text estimates or states that the effect of AI on aggregate productivity or GDP over roughly the next decade will be modest: a small effect on the level of output, not a change in the growth regime.
Earliest located, tier 1 Daron Acemoglu · The Simple Macroeconomics of AI · 2024-04-05 · wording attested 2025-01-06 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (1) Articulates · Earliest located · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-04-05) articulates G08. A task-based estimate using Hulten's theorem: the share of tasks exposed to AI times the average task-level cost saving. The GDP effect after an assumed investment response; under the investment assumptions of earlier automation the upper bound rises to 1.6–1.8%. Conditions: Task-based model; cost savings at the task level; existing estimates of exposure (about 4.6% of tasks affected within ten years) and of cost savings (15.4%). 2024-04-05 · wording attested 2025-01-06 Inspect claim & evidence ↗ Document page → G09 Estimates of explosive growth within a stated horizonCriterion. The text states its own estimate of whether or when explosive growth (as the text defines it) occurs in the actual future, with machine intelligence or automation among the stated drivers. Each record gives the holder, the estimate as stated (a number or words), the horizon, the text's definition and the conditions. Records are not converted to a common threshold, combined or ranked; a low estimate and a high estimate both articulate this proposition.
Earliest located, tier 1 Not assigned. Withheld on 2026-10-25 by maintainer decision. The earlier holder, Hanson’s “Economic Growth Given Machine Intelligence”, relays his separate 1998 projection. His “Long-Term Growth As A Sequence of Exponential Modes” (December 2000; first version September 1998) is an unchecked lead for the role.
Earliest located, tier 2 Holden Karnofsky · Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece") · 2020-06-15 · wording attested 2020-06-23 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (7) Articulates · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) articulates G09. The next sentence: the prediction "suggests that we not reject out of hand the predictions of this model of machine intelligence", and the following one links 2025 to brain-equivalent computer hardware. Estimate by Robin Hanson: median prediction of a transition to an economic doubling time of one to two years · horizon: around 2025 (measured from an undated file attested 2000) · definition: doubling time of world product of one to two years · conditional · conditions: "an empirical projection of historical trends in world economic growth"; the author adds that it "has large uncertainties" and links 2025 to brain-equivalent computer hardware in the next sentence. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Robin Hanson · Long-Term Growth As A Sequence of Exponential Modes Hanson, “Long-Term Growth As A Sequence of Exponential Modes” (2000-12) articulates G09. The next sentences point to falling computer hardware costs, predicted machine substitution for most human labour, and his own model of growth with robots. Links the forecast to machine substitution for labour, citing his own 1998 paper (G01-a). Conditions: "if one takes seriously the model of economic growth as a series of exponential growth modes, and if relative change parameters of a new transition are likely to be similar to such parameters describing old transitions". Estimate by Robin Hanson: "hard to escape the conclusion that the world economy could see a very dramatic change" · horizon: within the next century (from December 2000) · definition: a new growth mode with a doubling time of roughly two weeks or less · conditional · conditions: If the series-of-modes model is taken seriously and a new transition resembles past ones. 2000-12 · wording attested 2001-11-28 · claimed 1998-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates G09. Extrapolating the two positive tests puts growth above 20% a year at least a century away. Estimate by William D. Nordhaus: the economy "might plausibly cross the Singularity" in "100 years or more" · horizon: 100 years or more (from 2015) · definition: economic growth rate crossing 20% per year · conditional · conditions: Simple extrapolation of the two of seven tests that are positive; no probability and no duration given. 2015-09 · wording attested 2016-08-09 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Holden Karnofsky · Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece") Karnofsky, “Comment on "Modeling the Human Trajectory" ("Some personal reactions on this piece")” (2020-06-15) articulates G09. The same comment says he independently holds that transformative AI "could plausibly be developed in the next couple of decades". Estimate by Holden Karnofsky: "little reason to 'rule out' explosive growth" · horizon: the coming decades (from June 2020) · definition: not defined in the comment · conditions: Tentative agreement with Roodman's themes; the same comment says transformative AI "could plausibly be developed in the next couple of decades". 2020-06-15 · wording attested 2020-06-23 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Tom Davidson · Could Advanced AI Drive Explosive Economic Growth? Davidson, “Could Advanced AI Drive Explosive Economic Growth?” (2021-06-25) articulates G09. Explosive growth is defined in the report as more than 30% annual growth of gross world product. Follows: "On a different day my probability of explosive growth by 2100 might be as low as 15% or as high as 60%." Conditions: Explosive growth defined as >30% annual GWP growth; horizon 2100. Estimate by Tom Davidson: "plausible but far from certain"; >10% (Appendix G); "as low as 15% or as high as 60%" on different days · horizon: by 2100 (from June 2021) · definition: more than 30% annual growth of gross world product · conditions: Overall judgement across the report's models and objections. 2021-06-25 · wording attested 2021-07-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Matt Clancy; Tamay Besiroglu · The Great Inflection? A Debate About AI and Explosive Growth Clancy and Besiroglu, “The Great Inflection? A Debate About AI and Explosive Growth” (2023-06) articulates G09. Speakers: Tamay Besiroglu; Matt Clancy. Besiroglu's estimate conditional on AI automating the vast majority of tasks. In his closing turn he gives 25–75% "conditional on AGI". He separately puts the development of such AI this century at about 80%; he does not combine the figures. Clancy's estimate on the same condition, with "most of that probability clustered around the low end of explosive growth". Estimate by Tamay Besiroglu: 65%; at the close 25–75% "conditional on AGI" · horizon: not stated; conditional on the arrival of the AI described · definition: at least ten times the annual growth rate of the past century, sustained for at least a decade · conditional · conditions: AI that automates the vast majority of tasks currently performed by humans. Estimate by Matt Clancy: 10% to 20%, "depending on how I'm feeling" · horizon: not stated; conditional on deploying AI throughout the economy · definition: Besiroglu's definition (at least tenfold the past century's growth rate for at least a decade) · conditional · conditions: AI deployed throughout the economy. 2023-06 · wording attested 2023-12-23 · claimed 2023 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ege Erdil; Tamay Besiroglu · Explosive growth from AI automation: A review of the arguments Erdil and Besiroglu, “Explosive growth from AI automation: A review of the arguments” (2023-10-01) articulates G09. The authors' unconditional estimate. The same paragraph says high confidence "seems unwarranted". Conditions: Explosive growth defined in the paper as annual real GWP exceeding 130% of its maximum value over all previous years (about ten times growth rates typical of today's frontier economies). Estimate by Ege Erdil and Tamay Besiroglu: about even odds of widespread automation and subsequent explosive growth · horizon: by the end of this century (from October 2023) · definition: annual real gross world product exceeding 130% of its maximum over all previous years · conditions: The same paragraph says high confidence "seems unwarranted". 2023-10-01 · wording attested 2023-10-01 · claimed 2023-09-20 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (1) Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-04-05) bears on G09. Compares his ten-year estimate with larger forecasts. A footnote cites Davidson (2021) among those who expect "explosive" growth. Missing: Concerns the next ten years and current AI; does not assess explosive growth this century. 2024-04-05 · wording attested 2025-01-06 Inspect claim & evidence ↗ Document page → Headlines replaced Hanson’s undated “Economic Growth Given Machine Intelligence” (attested 2000-08-19), headlined as earliest located from 2026-10-10. Removed after the re-checks of 2026-10-25: the passage relays the author’s separate, self-cited projection, not an estimate the text makes itself. No tier-1 text holds the headline until the lead named under “How we searched” is checked (maintainer decision of 2026-10-25). Recorded 2026-10-25. Hanson’s “Long-Term Growth As A Sequence of Exponential Modes” (2000), headlined as earliest located in the first pass. Earlier text found by the audit: Hanson’s undated “Economic Growth Given Machine Intelligence”, attested 2000-08-19. Recorded 2026-10-10. M · Mainstream economics, automation and AI How economists and economics venues treated automation, computers and AI from 1952 to 2024: as ordinary technical change, as the automation of tasks, as a general-purpose technology or as cheaper prediction, and what they said about wages, policy, measurement and the field itself. M06 is split into M06a and M06b. M09 states the prior that mainstream economics took up machine intelligence only recently; it is tested against the dated record, not confirmed.
M01 Computers and AI as ordinary technical changeCriterion. A tier-1 text dated before 2015 that discusses computers, automation or AI and analyses them with the tools used for technical change in general (factor-augmenting or skill-biased change, capital deepening), with no separate category for machine cognition.
Earliest located, tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings · 1955-10-14 ↗
Earliest located tier-2 text that contests it Nils J. Nilsson · Artificial Intelligence, Employment, and Income · 1984 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (6) Articulates · Earliest located · Tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) articulates M01. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Under 'mechanization', the first of four principles of automation, Buckingham says mechanization sometimes substitutes machinery for muscle (the steam engine) and sometimes substitutes machinery for routine brainwork (the electronic computer). Both appear as cases of one general principle. Conditions: Restricted to brainwork 'at the lower, routine levels'. 1955-10-14 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) articulates M01. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse's first point: automation is 'only one inseparable part of the larger problem of technological progress', alongside jet propulsion, new metallurgy, plastics and atomic fission. He says it 'is only a way of increasing this varied and highly productive technology'. Conditions: The sentence continues: 'of trying to capture the values of higher productivity put within our reach by scientific progress'. 1955-10-28 Inspect claim & evidence ↗ Document page → Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) contests M01. After reviewing Simon's comparative-advantage argument and Leontief's input-output work, Nilsson argues that AI differs in kind from earlier labour-saving technologies, so the usual expansion-of-employment reasoning does not carry over. 1984 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Alan B. Krueger · How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984-89 Krueger, “How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984-89” (1991-10) articulates M01. The conclusion places the computer premium inside the skill-biased technological change explanation of 1980s wage-structure changes, against the competing trade explanation. 1991-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Timothy Bresnahan; Erik Brynjolfsson; Lorin Hitt · Information Technology, Workplace Organization and the Demand for Skilled Labor: Firm-Level Evidence Bresnahan, Brynjolfsson, Hitt, “Information Technology, Workplace Organization and the Demand for Skilled Labor: Firm-Level Evidence” (1999-05) articulates M01. The authors analyse information technology as skill-biased technical change working through complementary organisational change, not only substitution. 1999-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jeffrey D. Sachs; Laurence J. Kotlikoff · Smart Machines and Long-Term Misery Sachs and Kotlikoff, “Smart Machines and Long-Term Misery” (2012-12) articulates M01. 'Smart machines' enter the model as machinery whose efficiency parameter u rises, a factor-augmenting technical change in a nested CES production function with unskilled and skilled labour. Conditions: Machines and unskilled labour combine in a CES intermediate; the intermediate combines with skilled labour; u is machine efficiency. 2012-12 · wording attested 2013-01 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (16) Bears on · Tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) bears on M01. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Buckingham lists three things called automation in 1955: electronic computers processing data, automatic machines on assembly lines, and the hypothetical fully automatic enterprise. Missing: Describes what electronic computers can do; it does not by itself say whether the economist's analysis treats them as ordinary technical change. That framing is on p. 31 (mechanization principle). Conditions: Choice is limited to 'previously anticipated' courses of action under built-in criteria. 1955-10-14 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) bears on M01. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse says the hearings, from Ford's vice president to Vannevar Bush, traced automation as mechanization: new devices, with electronics applied to the control of mechanical and chemical processes. Commenting on Buckingham's four principles, Nourse argues that continuous process is old and that the real change came with electronic feedback control. Missing: (M01-03) Describes the computer's role as substitution of mechanical for mental work; no economic analysis of it as ordinary technical change in this sentence (see M01-05). (M01-04) Nourse distinguishes feedback control technologically; this does not dispute analysing automation as ordinary economic technical change. Conditions: (M01-03) Summarises the witnesses' account of 'automation as mechanization'. (M01-04) The newness is technological, in feedback control. He continues: 'It has its roots in mechanization, to be sure, but something new was added.'. 1955-10-28 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Frank Lynn · An Investigation of the Rate of Development and Diffusion of Technology in Our Modern Industrial Society Lynn, “An Investigation of the Rate of Development and Diffusion of Technology in Our Modern Industrial Society” (1966-02) bears on M01. Speaker: Frank Lynn, INTEC, Incorporated, Chicago (consultant; not an economist). Lynn measures incubation and diffusion periods across a set of innovation case studies; the computer is one case. Missing: Technological description inside a diffusion case study. The computer is one of several case-study innovations (with nuclear power, numerical control and others) analysed by the same time-lag method. No economic category. Conditions: The same section says the 1948 stored program and memory 'transformed the computer from a high-speed adding machine to its present status of “electronic brain.”' (the term is in quotation marks in the source). 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Edwin Mansfield · Technological Change: Measurement, Determinants, and Diffusion Mansfield, “Technological Change: Measurement, Determinants, and Diffusion” (1966-02) bears on M01. Speaker: Edwin Mansfield, University of Pennsylvania (economist), 'Technological Change: Measurement, Determinants, and Diffusion'. Commissioned survey of how economists measure technical change (partial and total productivity indexes, organisational and capital-embodied change). Part II, sec. 8 asks whether the rate has increased, quoting Diebold and DuBridge on automation. Missing: A general definition of technological change, not an analysis of computers or automation. Conditions: General definition; applies 'regardless of its source or form'. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Philip Sporn; Patrick E. Haggerty · Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I Sporn and Haggerty, “Comment by Mr. Sporn, joined by Mr. Haggerty, in Technology and the American Economy, Vol. I” (1966-02) bears on M01. Speaker: Philip Sporn (utility executive; Commission member, non-economist), joined by Patrick E. Haggerty. Sporn says the report under-emphasises technology's positive contributions. A parallel comment by Beirne, Hayes and Reuther (p. 6) says the report lacks urgency. Missing: A comment by two non-economist members, not the Commission's analysis. Conditions: Individual members' dissenting-emphasis comment, not Commission text. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) bears on M01. The Commission says the present is a time of rapid progress but that 'this is as far as we are prepared to go'. It concludes that no sharp break in the continuity of technical progress has occurred or is likely in the next decade. Missing: Agnostic on whether the computer is more revolutionary than earlier technologies; gives no positive analysis of it as ordinary technical change. Conditions: Agnostic rather than a denial. The chapter then measures technological change through output per man-hour and diffusion lags, the same measures for every innovation. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M01. Speaker: Wassily Leontief and Faye Duchin (economists). Puts the 1984 projections in a long-run perspective. Missing: A historical analogy that treats the 'electronic revolution' like mechanical power. It is not an analytical treatment of computers as ordinary technical change, though the report's method (technical coefficients in an input-output model) does treat them that way. Conditions: Preceded by: the industrial revolution of mechanical power continued 'over a period of some two hundred years'. 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Alan B. Krueger · How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984-89 Krueger, “How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984-89” (1991-10) bears on M01. Krueger frames computers as a measurable workplace technology and estimates a wage premium for computer users from CPS and High School and Beyond microdata. Missing: States the research question; the framing as technical change is in M01-13. 1991-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” (2013-09-17) bears on M01. The authors judge the routine/non-routine task framework inadequate for machine learning and mobile robotics and add engineering-bottleneck dimensions drawn from computer science. Missing: Neither clearly ordinary technical-change analysis (it revises the task model's premises about what computers can do) nor a separate category for machine cognition (computer capital remains an input at a falling price). 2013-09-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on M01. Nordhaus models AI-driven growth with a standard neoclassical growth model in which information capital receives capital-augmenting technical change; the outcome turns on the elasticity of substitution. Missing: Dated September 2015, after the working boundary; the paper also discusses superintelligence as a distinct prospect drawn from computer science. Conditions: Constant savings rate, closed economy, capital-augmenting technical change, CES production. 2015-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) bears on M01. Positions the task model against the "more commonly-used" factor-augmenting models, which the authors say do not produce these effects. Missing: Not a pre-2015 text; it characterises the prevailing factor-augmenting approach without naming the earlier texts. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) bears on M01. Places AI in a continuous history of automation from the steam engine to semiconductors, while noting AI may now reach non-routine cognitive tasks. Missing: Not a pre-2015 text; a 2017 statement of continuity between AI and earlier automation. 2017-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) bears on M01. Argues that the canonical factor-augmenting production function cannot represent automation's displacement of labour. Missing: Not a pre-2015 text; characterises the canonical F(AL, BK) approach without dating particular texts. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Modeling Automation Acemoglu and Restrepo, “Modeling Automation” (2018-05) bears on M01. Opens by rejecting the factor-augmenting treatment of automation, implying it was an available modelling convention that the authors replace with a task-based one. Missing: Not a pre-2015 text; it criticises the factor-augmenting approach rather than exemplifying it, and names no earlier texts. 2018-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019-03) bears on M01. The authors contrast their task framework with the canonical factor-augmenting production function, which they say lacks descriptive realism. Missing: A later text describing the canonical approach; not itself a pre-2015 text. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M01. The authors describe the standard approach as treating technology as making labour more productive across all tasks, which they reject as unrealistic. Missing: A later characterisation, not a pre-2015 text analysing computers as ordinary technical change. 2019-03 Inspect claim & evidence ↗ Document page → Headlines replaced Nourse’s statement of 1955, headlined as the earliest located text that contests it. Reclassified as bearing on it after the cross-family check. No tier-1 text now contests M01. Recorded 2026-10-12. M02 AI as the automation of tasksCriterion. A text that models or describes AI's economic effect as machines taking over tasks previously done by labour, within a task framework (tasks allocated between capital and labour).
Earliest located, tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? · 2013-09-17 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (10) Articulates · Earliest located · Tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” (2013-09-17) articulates M02. Extending the Autor-Levy-Murnane task model, the authors argue that machine learning lets computer capital substitute for labour in non-routine as well as routine tasks, limited only by perception/manipulation, creative and social-intelligence bottlenecks. Conditions: Computer capital supplied perfectly elastically at a falling price; workers reallocate by comparative advantage. 2013-09-17 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 David H. Autor · Why Are There Still So Many Jobs? The History and Future of Workplace Automation Autor, “Why Are There Still So Many Jobs? The History and Future of Workplace Automation” (2015) articulates M02. Autor describes automation as computers substituting for workers in routine, codifiable tasks while complementing workers in non-routine tasks. 2015 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) articulates M02. Introduces a task framework in which automation moves labour-performed tasks to capital while new tasks in which labour has comparative advantage are created; motivated by concerns about robotics and AI. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates M02. Frames AI within a task-based automation model (building on Zeira 1998 and Acemoglu and Restrepo) in which tasks move from labour to capital. 2017-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) articulates M02. Applies the task framework explicitly to AI and robotics, emphasising a displacement effect absent from factor-augmenting models. 2018-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment Acemoglu and Restrepo, “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment” (2018-06) articulates M02. The abstract sets the paper in a framework in which labour-performed tasks are automated while new labour-advantaged tasks are created, to examine concerns that labour will become redundant. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019) articulates M02. The published JEP abstract defines automation in a task framework as capital replacing labour in tasks. Conditions: Task-based framework with new tasks as counterweight. 2019 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) articulates M02. The first of four direct effects in the authors' task framework is capital substituting for labour in prediction tasks. Conditions: Task framework with prediction and decision tasks. 2019-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) articulates M02. The authors frame AI as a platform that can either automate labour's tasks or create new ones, and argue current development is biased toward automation. Conditions: Task framework (automation versus new tasks). 2019-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) articulates M02. Acemoglu models AI in a task framework where automation expands the set of tasks allocated to capital, alongside task complementarity, deepening and new tasks. Conditions: Task-based model with tasks allocated between capital and labour by comparative advantage. 2024-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (14) Bears on · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) bears on M02. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse locates the 'real change' in electronic devices that remove the need for human judgement in process control. Missing: Describes machines taking over the mental element of control work, but not within a framework that allocates tasks between capital and labour, and not as AI. Conditions: 'to considerable extent'; limited to control of continuous production processes. 1955-10-28 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Joseph P. Newhouse · Technological Change in Banking Newhouse, “Technological Change in Banking” (1966-02) bears on M02. Speaker: Joseph P. Newhouse, Harvard University (byline; also listed as a Commission staff analyst), 'Technological Change in Banking'. Newhouse finds that aggregate statistics say little; case studies show EDP eliminating routine clerical jobs, with adjustment largely through attrition given high turnover. Missing: Identifies the occupations most vulnerable to electronic data processing (EDP) through case studies. It is occupational, not a task-allocation model, and concerns EDP, not AI. Conditions: The sentence begins 'the most vulnerable occupations are' (OCR garbles the following comma) 'as might be expected'. Case examples follow: bookkeepers at one branch bank fell from 600 to 150 within eighteen months. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) bears on M02. Speaker: Howard R. Bowen, Commission chairman (economist), chairman's introduction (not part of the official report). Bowen's view of where technology should go: towards human, environmental and resource development, with new kinds of work and more leisure. Missing: Divides work into routine/mechanical (machines) and new human activities, but without a task-allocation model and not about AI. Conditions: Followed by: 'Human resources will be released and available for new activities beyond those that are required for mere subsistence.'. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M02. Speaker: Wassily Leontief and Faye Duchin (economists). Introduces the chapter's procedures for representing computer-based technologies in capital, input and labour coefficients. Projecting managerial labour coefficients under office automation. Overview before estimating robot stocks and displacement. Missing: (M02-04) Task language without a task-allocation model; the model works through input-output labour coefficients by occupation and sector. Computers, not AI. (M02-05) Software replacing decision-making tasks of managers, stated descriptively; no task-allocation framework, not AI. (M02-06) Robots taking over tasks previously done only by humans, in prose; the model represents this as labour displacement rates per robot in six production occupations, not a task framework. Not AI. Conditions: (M02-04) Labour savings 'of particular significance' for banking, insurance, legal services and government. (M02-05) 'certain' tasks. The passage adds that fewer managers will be needed as integrated office systems reduce clerical staff. (M02-06) Supported by the Robot Institute of America definition, quoted by the authors. 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Jeffrey D. Sachs; Laurence J. Kotlikoff · Smart Machines and Long-Term Misery Sachs and Kotlikoff, “Smart Machines and Long-Term Misery” (2012-12) bears on M02. The abstract describes the model's mechanism: smart machines substitute for unskilled labour and complement skilled labour. Missing: Substitution is between factors (machines and unskilled labour) in a CES function, not machines taking over tasks within a task framework. 2012-12 · wording attested 2013-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Modeling Automation Acemoglu and Restrepo, “Modeling Automation” (2018-05) bears on M02. Argues for modelling automation as machines taking over labour-performed tasks rather than as factor-augmenting technical change. Missing: Automation in general; AI is not named in the abstract. 2018-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019) bears on M02. The abstract distinguishes cases where automated prediction automates decisions from cases where it enhances human decisions. Missing: Presents automation as one of two outcomes rather than AI's economic effect. 2019 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) bears on M02. The authors argue AI's labour effect is ambiguous (substitution, decision automation, augmentation, new tasks) and caution against inferring it from factory-automation research. Missing: The conclusion cautions against extrapolating factory-automation results; it explicitly accepts direct capital-for-labour substitution in prediction tasks and therefore does not contest the criterion. 2019-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019-03) bears on M02. The paper's framework allocates tasks between capital and labour and defines automation as capital taking over labour's tasks. Missing: Automation in general; AI is not named in the quoted words. Conditions: Task-based model; also includes new tasks, factor-augmenting change. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M02. The authors characterise prevailing researcher and economist views of AI as automation, then argue it need not be used mainly that way. Missing: Describes others' framing; the authors themselves contest that automation must be AI's primary use. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) bears on M02. The editors treat automation as the third of three AI concepts and note that much empirical work on AI uses factory-robot data, while distinguishing automation from AI itself. Missing: Treats task automation as one possible consequence of AI, not as AI's economic effect. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on M02. Cowen characterises the recent literature as focused on capital-for-labour substitution and its domestic labour-market effects, then turns to neglected questions. Missing: Describes the literature's focus on substitution rather than itself modelling AI as task automation. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Emilio Calvano; Giacomo Calzolari; Vincenzo Denicolò; Sergio Pastorello · Artificial Intelligence, Algorithmic Pricing, and Collusion Calvano et al., “Artificial Intelligence, Algorithmic Pricing, and Collusion” (2020-10) bears on M02. The abstract motivates the study by algorithms replacing human pricing decisions. Missing: No task framework allocating tasks between capital and labour; focus is collusion, not labour. 2020-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyna Eloundou; Sam Manning; Pamela Mishkin; Daniel Rock · GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models Eloundou et al., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (2023-03-17) bears on M02. The authors deliberately do not classify exposure as automation versus augmentation. Missing: Exposure is not identified with machines taking over tasks. 2023-03-17 Inspect claim & evidence ↗ Document page → Headlines replaced Agrawal, Gans and Goldfarb (2019), headlined as the earliest located text that contests it. Reclassified as bearing on it after the cross-family check. No tier-1 text now contests M02. Recorded 2026-10-12. M03 AI as a general-purpose technology with lagging productivityCriterion. A text that calls AI a general-purpose technology and states that measured productivity gains lag adoption because of complementary investment (for example a J-curve or a productivity paradox).
Earliest located, tier 1 Erik Brynjolfsson; Daniel Rock; Chad Syverson · Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics · 2017-11 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (2) Articulates · Earliest located · Tier 1 Erik Brynjolfsson; Daniel Rock; Chad Syverson · Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics Brynjolfsson, Rock, Syverson, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics” (2017-11) articulates M03. Argues that implementation lags, rather than false hopes, mismeasurement or redistribution, are the likeliest main explanation of the gap between AI expectations and slowing measured productivity. Concludes that AI breakthroughs have not yet diffused and their complementary changes depress measured productivity at first; the same conclusion calls AI the core of "a new, economically important potential GPT". 2017-11 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Economic Policy for Artificial Intelligence Agrawal, Gans, Goldfarb, “Economic Policy for Artificial Intelligence” (2018-06) articulates M03. Following the preceding paragraph's point that GPT co-invention "is costly and slow", infers that AI's productivity impact will lag, citing Brynjolfsson, Rock and Syverson. 2018-06 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (7) Bears on · Tier 1 Erik Brynjolfsson; Daniel Rock; Chad Syverson · Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics Brynjolfsson, Rock, Syverson, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics” (2017-11) bears on M03. Formalises how unmeasured intangible AI-complementary investment first depresses and later inflates measured productivity growth. Missing: The J-curve mechanism for intangible capital in general; AI is not named in the quoted words (the paper applies it to AI, see M03-01). Conditions: Unmeasured intangible complementary investment; toy-economy illustration. 2017-11 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) bears on M03. One of three possibilities raised against the slow recent productivity record, alongside mismeasurement and a singularity-type discontinuity. Missing: States a possible delay but does not call AI a general-purpose technology here; offered as one of three alternatives. 2017-12 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Manuel Trajtenberg · AI as the Next GPT: A Political-Economy Perspective Trajtenberg, “AI as the Next GPT: A Political-Economy Perspective” (2018-01) bears on M03. Frames AI as the coming GPT and turns to policies to soften its displacement effects. Missing: Calls AI the new GPT but does not state that measured productivity gains lag adoption. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) bears on M03. The authors call AI a general-purpose technology because prediction is an input into decisions across many occupations. Missing: No statement that measured productivity gains lag adoption. 2019-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) bears on M03. The editors summarise Brynjolfsson, Rock and Syverson (chapter 1), who argue AI is a GPT and attribute low measured productivity growth mainly to implementation lags from missing complementary innovations. Missing: The GPT-plus-lag conjunction is attributed to chapter 1's authors; the introduction only reports it. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Jason Furman · Should We Be Reassured If Automation in the Future Looks Like Automation in the Past? Furman, “Should We Be Reassured If Automation in the Future Looks Like Automation in the Past?” (2019-05) bears on M03. Furman attributes AI's small macroeconomic impact to the newness of recent advances and notes that inventions like electricity added productivity in successive waves. Missing: Does not call AI a general-purpose technology or attribute the lag to complementary investment. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyna Eloundou; Sam Manning; Pamela Mishkin; Daniel Rock · GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models Eloundou et al., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (2023-03-17) bears on M03. The authors argue LLMs meet general-purpose-technology criteria of pervasiveness, improvement and complementary innovation. The authors note that co-invention requirements typical of general-purpose technologies may slow LLM diffusion. Missing: (M03-10) No statement that measured productivity gains lag adoption. (M03-11) Refers to slowed diffusion, not to measured productivity lagging adoption. 2023-03-17 Inspect claim & evidence ↗ Document page → M04 AI as cheaper predictionCriterion. A text that defines or treats AI's economic function as prediction and draws consequences from a fall in its cost (for example for the value of complements such as judgement or data).
Earliest located, tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · The Simple Economics of Machine Intelligence · 2016-11-17 · wording attested 2016-11-25 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (6) Articulates · Earliest located · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · The Simple Economics of Machine Intelligence Agrawal, Gans, Goldfarb, “The Simple Economics of Machine Intelligence” (2016-11-17) articulates M04. Writing as economists, the authors treat machine intelligence as a prediction technology whose economic effect centres on a fall in the cost of prediction, and conclude that human judgment, a complement, becomes more valuable. 2016-11-17 · wording attested 2016-11-25 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Prediction, Judgment and Complexity: A Theory of Decision Making and Artificial Intelligence Agrawal, Gans, Goldfarb, “Prediction, Judgment and Complexity: A Theory of Decision Making and Artificial Intelligence” (2018-01) articulates M04. Treats AI as prediction technology and models how better prediction changes the value of judgment, the process of determining payoffs. Makes the fall in the cost of prediction the premise of the analysis. Draws the consequence of better prediction for a complement, human judgment. Conditions: (M04-03) Stated as an assumption of the model. (M04-04) Single-agent decision model with a risky and a safe action; holds while judgment is not too difficult; in complex environments results depend on whether prediction enables automated decision-making. 2018-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Economic Policy for Artificial Intelligence Agrawal, Gans, Goldfarb, “Economic Policy for Artificial Intelligence” (2018-06) articulates M04. Recasts machine learning as a fall in the price of prediction and asks how policy shapes its diffusion and consequences. Defines prediction and summarises the authors' programme on consequences of cheaper prediction. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019) articulates M04. The published abstract identifies AI with prediction and frames its labour impact through prediction and decision tasks. 2019 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) articulates M04. The authors identify current AI with prediction and draw labour-market consequences from cheaper prediction for complementary decision tasks. The authors set prediction as a perfect complement to decisions, so falling prediction costs change returns to labour and capital in decision tasks. Conditions: (M04-08) AI equated with machine-learning prediction; decisions and judgement as complements. (M04-09) Prediction and decision tasks as perfect complements. 2019-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) articulates M04. The editors summarise their own chapter, which models AI as a fall in the cost of prediction and draws organisational consequences, including complementarity with human judgement. Conditions: AI identified with machine-learning prediction; judgement treated as a complement. 2019-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (1) Bears on · Tier 1 Sendhil Mullainathan; Jann Spiess · Machine Learning: An Applied Econometric Approach Mullainathan and Spiess, “Machine Learning: An Applied Econometric Approach” (2017) bears on M04. Mullainathan and Spiess characterise machine learning as a prediction tool distinct from econometric parameter estimation, for use in empirical economics. Missing: Treats prediction as what ML does methodologically, not as AI's economic function, and draws no consequences from a fall in its cost. 2017 Inspect claim & evidence ↗ Document page → Headlines replaced Agrawal, Gans and Goldfarb’s working paper (January 2018), headlined as earliest located. Earlier text found by the re-checks of 2026-10-25: the same authors’ essay in the Harvard Business Review (17 November 2016, attested 2016-11-25), an economists’ magazine essay and so tier 1. Recorded 2026-10-25. M05 Machines cheaper than labour at every taskCriterion. A text that states or models the possibility that machines perform all or nearly all tasks, including tasks created later, at lower cost than human labour. Texts that assume new tasks in which labour is cheaper keep appearing bear on it without meeting it.
Earliest located, tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment · 2016-05 · wording attested 2017-06 ↗
Earliest located, tier 2 Nils J. Nilsson · Artificial Intelligence, Employment, and Income · 1984 ↗
Earliest located tier-1 text that contests it Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings · 1955-10-14 ↗
Earliest located tier-2 text that contests it National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I · 1966-02 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (14) Contests · Earliest located · Tier 1 · Engages a premise Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) contests M05. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Concludes his assessment of the employment criterion before turning to stability. Conditions: The next sentence qualifies it: there is 'no automatic regulator in the economic system that guarantees full employment', and the gains need continued expansion. He likens the economy to 'a rocket which must continue to accelerate or else fall from the sky'. 1955-10-14 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · States the contrary Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) contests M05. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse contrasts the 'preponderant attitude of business executives' with labour officials' concerns and goes on to side with labour's call for study (p. 621). Conditions: He adds at once that 'we need to study the fine print', given 'severe short-run dislocations, local catastrophes, and painful readjustments'. 1955-10-28 Inspect claim & evidence ↗ Document page → Contests · Tier 2 · States the contrary Paul Armer · Computer Aspects of Technological Change (selection) Armer, “Computer Aspects of Technological Change (selection)” (1966) contests M05. Speaker: Paul Armer, The RAND Corporation (non-economist). Armer combines technological radicalism (computers able in time to do any information-processing task) with economic conservatism. Conditions: He acknowledges that computers and automation cause serious problems needing economic, social and political adjustment. 1966 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · Engages a premise Sar A. Levitan · New Directions in Aid of the Poor (selection) Levitan, “New Directions in Aid of the Poor (selection)” (1966) contests M05. Speaker: Sar A. Levitan, The W. E. Upjohn Institute for Employment Research (economist), selection 'New Directions in Aid of the Poor'. Levitan's priorities for anti-poverty policy include work relief and public service jobs such as school and health aides. Conditions: Made in support of government-supported (not make-work) job creation for those who cannot find private employment. 1966 Inspect claim & evidence ↗ Document page → Contests · Earliest located in tier 2 · Tier 2 · Engages a premise National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) contests M05. Opening of the Commission's report, signed by all 14 members, including the economists Bowen and Solow. Concludes the Commission's opening argument that persistent post-Korea unemployment came from inadequate demand, not accelerated technology. Conditions: (M05-11) Also dissents from 'the other extreme view of complacency'. The ellipsis omits '— or at least the United States —'. (M05-12) Followed by: 'It is the continuous obligation of economic policy to match increases in productive potential with increases in purchasing power and demand.'. 1966-02 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) articulates M05. Applying a 'law of reduction', Nilsson argues that producers will replace human labour with cheaper AI in every activity, including the service and white-collar work to which labour has moved, and that too few nonautomatable services remain to base an economy on. Conditions: Assumes AI becomes cheaper than human labour across activities and that demand is finite. 1984 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · States the contrary David H. Autor · Why Are There Still So Many Jobs? The History and Future of Workplace Automation Autor, “Why Are There Still So Many Jobs? The History and Future of Workplace Automation” (2015) contests M05. Reflecting on AI and robotics, Autor expects workers to keep a comparative advantage in problem-solving, adaptability and creativity, against commentators who 'overstate the extent of machine substitution'. 2015 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 · Model scenario Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) articulates M05. States the full-automation case of the model, in which automation keeps pace with task creation so that labour becomes redundant; the authors immediately contrast it with the stable balanced-growth case on which they focus. Conditions: Full model with directed research toward automation versus new tasks; holds when the long-run rental rate of capital is very low relative to labour (in the formal results, a sufficiently low discount rate), giving a balanced growth path in which all tasks are produced with capital. Otherwise an interior, stable balanced growth path. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates M05. Models the case in which A.I. replaces people in all tasks to show how growth could explode; the authors then list objections (automation limits, search limits, Baumol bottlenecks). Conditions: Thought experiments: all tasks in goods production automated (AK growth) or in idea production automated (with φ > 0, a finite-time singularity). Automation is exogenous; the costs of machines versus labour are not modelled as a comparison of task-level prices. 2017-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates M05. Describes a transition in which falling machine costs make machines a cheaper substitute for all human labour; with frictionless capital adjustment the human wage may be unchanged while the labour share goes to zero. Conditions: Machine labour modelled as a perfect substitute for human labour; the cost of producing machines falls over time. The passage summarises the authors' companion paper (Korinek and Stiglitz 2017). 2017-12 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · States the contrary Manuel Trajtenberg · AI as the Next GPT: A Political-Economy Perspective Trajtenberg, “AI as the Next GPT: A Political-Economy Perspective” (2018-01) contests M05. Concludes, from history, that human obsolescence is not to be presumed, while expecting many occupations to vanish. 2018-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Daron Acemoglu; Pascual Restrepo · The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment Acemoglu and Restrepo, “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment” (2018-06) articulates M05. A conditional model result: under a low relative cost of capital, all tasks (including those created later) end up automated; the abstract presents the alternative stable balanced-growth case as the other outcome. Conditions: Full model with endogenous capital accumulation and directed research toward automation and new tasks; full automation only if the long-run rental rate of capital relative to the wage is sufficiently low. Otherwise a stable balanced growth path in which automation and new tasks go together. 2018-06 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · States the contrary Jason Furman · Should We Be Reassured If Automation in the Future Looks Like Automation in the Past? Furman, “Should We Be Reassured If Automation in the Future Looks Like Automation in the Past?” (2019-05) contests M05. Furman rejects mass unemployability and relocates the concern to prolonged transitions as displaced workers re-match to new jobs. 2019-05 Inspect claim & evidence ↗ Document page → Contests · Tier 1 · States the contrary Betsey Stevenson · Artificial Intelligence, Income, Employment, and Meaning Stevenson, “Artificial Intelligence, Income, Employment, and Meaning” (2019-05) contests M05. Stevenson reconciles divided expert views by distinguishing a long run in which new jobs arise from a short run of disruption. 2019-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (21) Bears on · Tier 1 · Relay of another author Wassily Leontief · Machines and Man Leontief, “Machines and Man” (1952-09) bears on M05. The standfirst of an article by the economist Wassily Leontief on automatic control systems asks whether they will cause widespread technological unemployment. Missing: Poses the question only; the article text, where Leontief answers it, was not retrieved. 1952-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 · Relay of another author Joint Committee on the Economic Report, Subcommittee on Economic Stabilization · Automation and Technological Change: Hearings (statements by non-economists) Subcommittee on Economic Stabilization of the Joint Committee on the Economic Report, “Automation and Technological Change: Hearings (statements by non-economists)” (1955-10) bears on M05. Speaker: UAW-CIO report (Economic and Collective Bargaining Conference, 12 November 1954), submitted for the record by Walter P. Reuther, president, CIO (non-economist witness). The UAW report collects predictions of rapid automation to argue for an expanding economy built on broad purchasing power. Missing: A relayed prediction of a fivefold labour saving, not machines performing all tasks. The economist is unnamed. Conditions: Attributed to 'another economist'. It is conditional on the preceding claim that, if expenditure on automating industrial facilities kept rising, 'American industry may be fully automated within a decade'. That claim rests on figures from 'a prominent Harvard economist', also unnamed. 1955-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) bears on M05. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Buckingham says this picture 'has charged the imagination of some and struck terror in the hearts of others', and that it too is called automation. He then complains of a 'flood of verbiage' lacking quantitative evidence. Missing: A reported speculation, not Buckingham's own claim; it concerns one enterprise working without labour, not machines performing every task, including new ones, more cheaply than people. Conditions: Attributed to 'scientists, computer manufacturers, and science fiction writers', who have shown it 'hypothetically at least'. 'chain or products' is as printed in the OCR. 1955-10-14 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Paul Armer · Computer Aspects of Technological Change (selection) Armer, “Computer Aspects of Technological Change (selection)” (1966) bears on M05. Speakers: Paul Armer, The RAND Corporation (computer specialist; not an economist), in a selection from a study prepared for the Commission; Paul Armer, The RAND Corporation, relaying Herbert A. Simon. Armer adopts Herbert A. Simon's distinction (The Shape of Automation, 1965) between technological and economic radicalism. Armer cites Simon, The Shape of Automation (New York: Harper & Row, 1965), as the source of the technological/economic radical-conservative distinction. Missing: (M05-05) A technological claim restricted to information-processing tasks. The author explicitly rejects the economic consequence (see the next record) and says nothing about relative cost or new tasks. (M05-06) Armer's paraphrase of Simon's technological forecast, not Simon's text. Armer reports Simon as an 'economic conservative', which would not meet M05's economic condition. Conditions: (M05-05) 'in time'. He says many remaining problems 'are very difficult' and calls himself a technological radical 'somewhat closer to the center' than Simon. (M05-06) Simon's view as characterised by Armer. 1966 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 · Relay of another author National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) bears on M05. Speaker: Howard R. Bowen, Commission chairman (economist), in a chairman's introduction stated not to be part of the official report. Bowen lists concerns about technology. The one 'perhaps most responsible for the establishment of the Commission' is the belief that technological change is a major source of unemployment. Missing: A reported fear, not endorsed. It does not speak of new tasks or relative cost. Conditions: The ellipsis omits 'technological change would in the near future not only cause increasing unemployment, but that'. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 · Relay of another author Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) bears on M05. Nilsson treats the eventual performance by AI of the new jobs that AI itself creates as a matter of time, alongside his argument (pdf p. 6) that ever-cheaper AI 'labor' will replace human labour until its cost share is insignificant. Missing: An objection Nilsson relays ("some might say") before answering it, not his own claim. 1984 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M05. Speaker: Wassily Leontief and Faye Duchin (economists). Concluding remarks on the scenario results. Missing: Agnostic on technological unemployment; says nothing of machines performing all tasks. Conditions: Scenarios are limited to computer-based automation and exclude other changes (agriculture, materials substitution) and 'any major break-throughs in computer technology'. 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” (2013-09-17) bears on M05. The authors claim near-universal technical automatability beyond the engineering bottlenecks, but expect workers to move into tasks requiring creative and social intelligence (pdf p. 45). Missing: Bottleneck tasks (creative and social intelligence) are assumed to remain with labour; no claim about cost of machines in all tasks or in newly created tasks. 2013-09-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Seth G. Benzell; Laurence J. Kotlikoff; Guillermo LaGarda; Jeffrey D. Sachs · Robots Are Us: Some Economics of Human Replacement Benzell et al., “Robots Are Us: Some Economics of Human Replacement” (2015-02) bears on M05. The model splits output into an automatable good produced by capital and code and a non-automatable service ('prayer') produced only by labour; the introduction (pdf p. 3) says it is 'hard to think of hitherto human tasks that smart machines can't already or won't soon do'. Missing: The model keeps a non-automatable labour-only sector, so machines do not perform all tasks; new tasks are not modelled. Conditions: Corn produced with capital and code; prayer produced with high- and low-tech labour only. 2015-02 · wording attested 2018-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on M05. In the accelerationist case of the model, information capital substitutes so well for labour that all income eventually accrues to capital owners, the reverse of Keynes's 'euthanasia of the rentier'. Nordhaus states that the Singularity would require AI to 'encompass all human activities' and notes that non-routine tasks change in response to AI, leaving immunity of other tasks 'an open question'. Nordhaus quotes Simon's 1965 statement of technological radicalism and contrasts it with Simon's moderate productivity expectation. Nordhaus reports that Simon, despite his capability forecast, did not expect accelerating productivity growth. Missing: (M05-18) A capital share tending to one does not imply every present and future task is performed more cheaply by machines; the continuation explicitly retains well-paid workers. (M05-19) Raises the role of evolving (new) tasks but neither asserts nor rules out machines performing them more cheaply; treats it as an empirical question of substitutability. (M05-41) A relayed capability forecast by Simon (1965), not a claim about cost relative to labour; in the same passage Nordhaus reports that Simon expected only a "continuing, but not greatly accelerated, rise in productivity". (M05-42) A relayed productivity expectation; it does not address cost relative to labour in all tasks. Conditions: Elasticity of substitution between information capital and labour bounded above one; rapid capital-augmenting technical change. 2015-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) bears on M05. Replies to Leontief's horse analogy (quoted earlier in the paragraph and attributed to him) by positing labour's comparative advantage in new tasks. Missing: Assumes that newly created tasks are ones in which labour has a comparative advantage; the passage answers the redundancy concern rather than modelling machines performing new tasks. Conditions: Model assumption: labour has comparative advantage in new, more complex tasks; employment and labour share stable if new-task creation continues. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) bears on M05. Treats complete automation as an open, debated question and names incomplete automation as a bottleneck to explosive growth. Missing: Leaves open whether machines can perform all essential tasks; neither asserts nor denies the possibility. Conditions: Conditional: if some essential tasks remain non-automated, the labour-free singularities do not occur. 2017-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) bears on M05. Identifies new labour-intensive tasks as the main force offsetting automation and keeping growth balanced. Missing: Assumes new tasks in which labour has comparative advantage; does not model machines performing new tasks. Conditions: New tasks are defined as ones where labour has comparative advantage relative to machines. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Prediction, Judgment and Complexity: A Theory of Decision Making and Artificial Intelligence Agrawal, Gans, Goldfarb, “Prediction, Judgment and Complexity: A Theory of Decision Making and Artificial Intelligence” (2018-01) bears on M05. Notes that existing automation models (here Acemoglu and Restrepo 2017) rely on tasks that only labour can perform, and offers a prediction-based approach instead. Missing: Identifies and questions the new-tasks assumption in existing models but does not itself model machines performing all tasks. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Manuel Trajtenberg · AI as the Next GPT: A Political-Economy Perspective Trajtenberg, “AI as the Next GPT: A Political-Economy Perspective” (2018-01) bears on M05. Summarises a prevailing view that AI will perform most new tasks too, so net displacement will exceed past GPTs. Missing: Reports what the author calls "the consensual view" and "the prevailing view" at the conference; the author does not adopt it. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) bears on M05. The authors deny that current machine learning amounts to general intelligence able to substitute for all of human cognition. Missing: Describes current machine learning's scope; does not address whether machines could eventually perform all tasks. 2019-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019-03) bears on M05. The authors acknowledge new tasks in which capital is advantaged but restrict their analysis to labour-intensive new tasks. The authors reject both imminent end-of-work claims and automatic labour gains, making the future depend on the balance of automation and new tasks. Missing: (M05-33) Recognises capital-advantaged new tasks but excludes them; does not consider machines performing all new tasks more cheaply. (M05-34) Addresses whether the end of work is imminent, not whether machines could eventually do all tasks including new ones. Conditions: Model assumes labour has comparative and absolute advantage in higher-indexed tasks. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M05. The authors warn that automation without new tasks would depress labour demand while denying an imminent end of work. Missing: Considers automation without new tasks, not machines also performing new tasks more cheaply; rejects near-term end of work. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) bears on M05. The editors name artificial general intelligence that equals or exceeds human intelligence as one of three AI concepts in the volume and argue its economic consequences are worth exploring regardless of timing. Missing: Raises machines surpassing human intelligence as a topic; does not state or model machines performing all tasks, including new ones, more cheaply than labour. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on M05. Cowen imagines a world where labour scarcity matters little and software generates and implements new products, so land and energy become the binding inputs. Missing: Imagines software generating new products and labour mattering little, but does not state machines perform all tasks more cheaply than labour. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on M05. Acemoglu lists a revolutionary, near-singularity scenario among four possible macro effects, says it cannot be ruled out, and excludes it from his analysis. Acemoglu argues that better wage and productivity outcomes depend on AI creating new tasks for workers, which he says is feasible in theory but not the industry's current focus. Missing: (M05-39) Acknowledges a transformative scenario as possible but does not state or model machines doing all tasks, including new ones, more cheaply. (M05-40) Presupposes that new tasks in which labour is productive can be created; does not consider machines outcompeting labour in them. 2024-05 Inspect claim & evidence ↗ Document page → Headlines replaced Nordhaus (2015), headlined as earliest located. Reclassified as bearing on it after the cross-family check. Acemoglu and Restrepo’s working paper, dated by its checked revision of June 2017, is now earliest located, as a model scenario. Recorded 2026-10-12. M06a Automation lowers the labour share, wages or employmentCriterion. A text that states or models that AI or automation lowers the labour share, wages or employment for some or all workers.
Earliest located, tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings · 1955-10-14 ↗
Earliest located, tier 2 John Diebold · A Condensation of the Book Automation, in Automation and Technological Change: Hearings · 1955-10-14 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (23) Articulates · Earliest located · Tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) articulates M06a. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Buckingham classifies eight direct effects of automation, starting with abolished production jobs and fewer new jobs that need higher skills. Citing a Chicago dissertation by David G. Osborn (12 automation cases, average 63.4 per cent reduction in employee requirements) and industry figures, Buckingham argues that displacement takes the form of reduced hiring and 'hidden unemployment of downgrading'. Conditions: (M06a-01) These are listed as direct effects. He goes on to argue that long-run mass unemployment is not to be feared (p. 35). (M06a-02) He says there have been no mass layoffs because automation has been slow enough for normal turnover to disguise displacement. 1955-10-14 Inspect claim & evidence ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 John Diebold · A Condensation of the Book Automation, in Automation and Technological Change: Hearings Diebold, “A Condensation of the Book Automation, in Automation and Technological Change: Hearings” (1955-10-14) articulates M06a. Under “Labor resistance” the condensation says that when automation makes a specific human skill obsolete it works a hardship on the individual worker who is displaced. 1955-10-14 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) articulates M06a. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse asks whether automation will disturb the price-and-income relations behind the postwar prosperity, and lists its effects on wages, capital, costs, prices, profits and taxes. Conditions: Mixed effects: the sentence continues that wage rates will be 'upgraded here and downgraded there'. 1955-10-28 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) articulates M06a. The Commission separates aggregate unemployment (demand) from displacement of particular workers (technology among other changes). Conditions: Distinguished from the general level of unemployment, which the Commission attributes to inadequate aggregate demand. 1966-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) articulates M06a. From the cost-reduction argument and the identity between labour costs and labour income, Nilsson concludes that labour's share of income will fall to a small fraction. 1984 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) articulates M06a. Speaker: Wassily Leontief and Faye Duchin (economists). The authors find that automation's effects are specific to types of work, with professionals rising and clerical workers falling sharply. Conditions: Under Scenario S3, by 2000, as a share of labour requirements. Professionals rise to nearly 20 per cent. 1984-04 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jeffrey D. Sachs; Laurence J. Kotlikoff · Smart Machines and Long-Term Misery Sachs and Kotlikoff, “Smart Machines and Long-Term Misery” (2012-12) articulates M06a. The authors argue, against the Cobb-Douglas presumption that technical change raises all factor returns, that better machines can lower unskilled wages. Conditions: Elasticity of substitution between machines and unskilled labour high relative to that between the intermediate and skilled labour. 2012-12 · wording attested 2013-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Seth G. Benzell; Laurence J. Kotlikoff; Guillermo LaGarda; Jeffrey D. Sachs · Robots Are Us: Some Economics of Human Replacement Benzell et al., “Robots Are Us: Some Economics of Human Replacement” (2015-02) articulates M06a. The authors state that their model predicts a long-run fall in labour's share as code accumulates, and point to the measured decline in Section 6. Conditions: Decline in the code depreciation rate in an OLG model with rival, excludable code (baseline case). 2015-02 · wording attested 2018-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates M06a. Extending the growth model to heterogeneous labour, Nordhaus shows that if unskilled labour is a perfect substitute for information capital its wage falls with capital prices; with a reservation wage its employment approaches zero. Conditions: Unskilled labour a perfect substitute for information capital; skilled labour highly but imperfectly substitutable. 2015-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) articulates M06a. Model result that automation lowers employment and the labour share and may lower wages. Conditions: Static model, fixed capital, exogenous technology; new tasks have the opposite effects; in the dynamic model the labour share still falls with automation while long-run real wages rise. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates M06a. Concludes that without redistribution, AI and other worker-replacing change can lower both labour's share and workers' absolute welfare. Conditions: Absence of redistribution or full insurance against innovation risk. 2017-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) articulates M06a. Even with offsetting forces, automation lowers the labour share. Conditions: Countervailing productivity, capital-accumulation and deepening effects present; absent new labour-intensive tasks. 2018-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Modeling Automation Acemoglu and Restrepo, “Modeling Automation” (2018-05) articulates M06a. Contrasts task automation with factor-augmenting change: the former always lowers the labour share. Conditions: Task-based model; the abstract adds that automation "can reduce the equilibrium wage (for realistic parameter values)". 2018-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment Acemoglu and Restrepo, “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment” (2018-06) articulates M06a. In the static model automation lowers employment and the labour share and may reduce wages, while new tasks push the other way. Conditions: Static version with fixed capital and exogenous technology; creation of new tasks has the opposite effects. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Economic Policy for Artificial Intelligence Agrawal, Gans, Goldfarb, “Economic Policy for Artificial Intelligence” (2018-06) articulates M06a. Gives a rising capital share as one of two reasons AI might increase inequality, with skill bias the other. Conditions: Conditional on AI being a new, efficient form of capital. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jason Furman; Robert Seamans · AI and the Economy Furman and Seamans, “AI and the Economy” (2018-06) articulates M06a. Opens the labour-market policy section by noting that AI may worsen declining participation and rising inequality. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019) articulates M06a. The abstract states that automation always lowers the labour share in value added. Conditions: Displacement effect, absent offsetting new tasks. 2019 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua S. Gans; Avi Goldfarb · Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction Agrawal, Gans, Goldfarb, “Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction” (2019-02) articulates M06a. The authors argue that workers whose core skill is prediction may lose, while those whose core skill complements prediction may gain. 2019-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019-03) articulates M06a. Within the task framework, automation's displacement effect lowers the labour share and can lower labour demand, offset by new tasks. Conditions: Holding new-task creation fixed; reinstatement from new tasks counterbalances. 2019-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) articulates M06a. The authors attribute stagnating labour demand, a falling labour share and rising inequality to technological change biased toward automation. 2019-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jason Furman · Should We Be Reassured If Automation in the Future Looks Like Automation in the Past? Furman, “Should We Be Reassured If Automation in the Future Looks Like Automation in the Past?” (2019-05) articulates M06a. Furman argues that, like past innovations, AI may further reduce participation and employment, particularly for less-skilled men, through slow re-matching of displaced workers. 2019-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Betsey Stevenson · Artificial Intelligence, Income, Employment, and Meaning Stevenson, “Artificial Intelligence, Income, Employment, and Meaning” (2019-05) articulates M06a. Stevenson says the real uncertainty is short-run disruption, with some workers harmed by falling demand for their skills. 2019-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) articulates M06a. Using Acemoglu and Restrepo (2022) general-equilibrium estimates, Acemoglu finds AI's wage effects more evenly spread than earlier automation but negative for low-education women and widening the capital–labour income gap. Conditions: Exposure measures from Eloundou et al. with easy/hard task distinction; propagation matrix from 1980–2016. 2024-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (3) Bears on · Tier 1 · Relay of another author National Commission on Technology, Automation, and Economic Progress · The Employment Impact of Technological Change. Appendix Volume II, Technology and the American Economy National Commission on Technology, Automation, and Economic Progress, “The Employment Impact of Technological Change. Appendix Volume II, Technology and the American Economy” (1966-02) bears on M06a. Speaker: Garth L. Mangum, Executive Secretary (Preface), summarising James R. Bright's study. The preface summarises the volume's 11 studies. Missing: A finding on skill content, not on wages, labour share or employment. It is a summary, and Bright's study (Part 3, p. 203) is not in the retrieved OCR. Conditions: Based on case studies of 'actual installations of automatic equipment'. The preface adds that Horowitz and Herrnstadt found 'no significant changes in skill requirements in general' between the 1949 and 1960 DOT editions. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) bears on M06a. Argues that Baumol's cost disease can keep the labour share substantial even as automation advances. Missing: A model in which the labour share stays substantial despite nearly complete automation; it does not state a fall for some workers, though the paper also notes a possible prolonged high-capital-share period. Conditions: Elasticity of substitution across tasks below one; automation exogenous; sufficient conditions for balanced growth with a constant capital share well below 100%. 2017-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on M06a. Cowen speculates that AI could harm poorer countries by removing wage-differential reasons to locate production there, cutting off industrialisation. Missing: Concerns poorer countries' path to industrialisation, not wages, employment or the labour share as such. Conditions: AI and other technologies progressing so firms hardly employ labour. 2019-05 Inspect claim & evidence ↗ Document page → Headlines replaced The report of the National Commission on Technology, Automation, and Economic Progress (1966), headlined as the earliest located tier-2 text. Earlier text found by the re-checks of 2026-10-25: the condensation of Diebold’s book Automation printed with his statement of 14 October 1955 in the congressional hearings. Recorded 2026-10-25. M06b Compensating policy for AI’s distributional effectsCriterion. A text that argues for redistribution, taxation or another compensating policy in response to distributional effects of AI or automation.
Earliest located, tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings · 1955-10-28 ↗
Earliest located, tier 2 John Diebold · A Condensation of the Book Automation, in Automation and Technological Change: Hearings · 1955-10-14 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (13) Articulates · Earliest located in tier 2 · Tier 2 John Diebold · A Condensation of the Book Automation, in Automation and Technological Change: Hearings Diebold, “A Condensation of the Book Automation, in Automation and Technological Change: Hearings” (1955-10-14) articulates M06b. The condensation says that responsibility for retraining and restoring the earning power of displaced workers lies with management as well as with the worker, and that accepting a fair share of it will ease the labour problem that automation brings. Conditions: Presented as a growing realization among businessmen; the next sentence says that accepting a fair share of responsibility for a cooperative solution will ease the labour problem. 1955-10-14 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Walter P. Reuther · Statement (president, Congress of Industrial Organizations), in Automation and Technological Change: Hearings Reuther, “Statement (president, Congress of Industrial Organizations), in Automation and Technological Change: Hearings” (1955-10-17) articulates M06b. Reuther urges the subcommittee to consider earlier Social Security payments for workers forced into retirement before 65 because technological changes have taken their jobs and their age keeps them from finding other work. 1955-10-17 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) articulates M06b. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse suggests that adoption might come closer to the optimum if companies had to provide for workers' costs of change, and lists severance pay and reimbursement for time lost or skills scrapped as devices to explore. Conditions: In the same sentence he says the individual need not be 'carried around on a pillow' and that government should not be 'the benevolent grandmother'. He calls for study of 'properly distributing the burdens'. 1955-10-28 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) articulates M06b. One of the Commission's numbered recommendations, alongside public service employment ('employer of last resort'), compensatory education and a computerised job-matching system. Conditions: 'That floor should include both improvements in wage-related benefits and a broader system of income maintenance for those families unable to provide for themselves.'. 1966-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) articulates M06b. Rejecting 'Luddite' obstruction and make-work jobs, Nilsson favours decoupling income from work, reviewing transfer payments, Kelso and Adler's broadened capital ownership and Albus's National Mutual Fund. 1984 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jeffrey D. Sachs; Laurence J. Kotlikoff · Smart Machines and Long-Term Misery Sachs and Kotlikoff, “Smart Machines and Long-Term Misery” (2012-12) articulates M06b. Rejecting 'smashing the machines', the authors propose intergenerational tax-and-transfer policy (in the model, a wealth levy on the old funding transfers to the young) to make all generations gain. Conditions: Model illustration: unanticipated permanent rise in u from 1 to 10; government levies part of the machine stock and transfers its returns. 2012-12 · wording attested 2013-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Seth G. Benzell; Laurence J. Kotlikoff; Guillermo LaGarda; Jeffrey D. Sachs · Robots Are Us: Some Economics of Human Replacement Benzell et al., “Robots Are Us: Some Economics of Human Replacement” (2015-02) articulates M06b. The conclusion argues that redistributive fiscal policy (taxing high-skilled workers who gain and investing the proceeds) can prevent immiseration, while blunt policies such as restricting labour supply or weakening IP can backfire. Conditions: Immiseration requires low saving propensity and low substitutability between capital and code. 2015-02 · wording attested 2018-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates M06b. Ties the case for redistribution to the distributional losses of worker-replacing technological change. Argues that policy, including non-distortionary taxation, can compensate those who lose from AI, even in a singularity case. Names patent length and capital taxation as second-best redistribution instruments for losses from worker-replacing machines. Conditions: (M06b-11) "Such intervention" refers to full insurance or "the right form of redistribution" named in the preceding sentence. (M06b-12) Conditions stated in the body: e.g., a fixed, irreproducible factor that captures the gains and can be taxed without distortion; or low supply elasticity of capital for capital taxation. (M06b-13) Lump-sum transfers unavailable; capital supply elasticity sufficiently low. 2017-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Manuel Trajtenberg · AI as the Next GPT: A Political-Economy Perspective Trajtenberg, “AI as the Next GPT: A Political-Economy Perspective” (2018-01) articulates M06b. Calls for policies (education, professionalising care work, steering technical change) to ameliorate AI's detrimental effects on displaced workers. Argues that mass displacement and ageing may require governments to manage transitions, not merely compensate losers. Conditions: Conditional on AI triggering mass displacement alongside population ageing. 2018-01 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Economic Policy for Artificial Intelligence Agrawal, Gans, Goldfarb, “Economic Policy for Artificial Intelligence” (2018-06) articulates M06b. Concludes that AI's consequences for jobs, inequality and competition call for education policy, the safety net and antitrust. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jason Furman; Robert Seamans · AI and the Economy Furman and Seamans, “AI and the Economy” (2018-06) articulates M06b. Weighs UBI against EITC expansion and wage subsidies as responses to AI-related labour disruption, favouring the latter. 2018-06 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) articulates M06b. Cowen suggests that if AI makes outputs very cheap, wealth-based transfers could complement education and training as policy responses to automation. Conditions: Very high productivity and very low output prices from AI. 2019-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Jason Furman · Should We Be Reassured If Automation in the Future Looks Like Automation in the Past? Furman, “Should We Be Reassured If Automation in the Future Looks Like Automation in the Past?” (2019-05) articulates M06b. Furman lists policy responses to AI's labour-market effects, including training and a more progressive tax system, while rejecting replacement of the safety net with a UBI. 2019-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (12) Bears on · Tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) bears on M06b. Speaker: Walter S. Buckingham Jr., associate professor, Georgia Institute of Technology (economist). Buckingham argues that automation serves consumer satisfaction, efficiency and progressiveness, but that mass purchasing power must be maintained. He also notes that automation may not create the waves of secondary investment that the automobile did. Missing: States a distributional condition (sharing through wages and prices) but proposes no redistribution, tax or compensating public policy. Conditions: Framed as a 'danger spot' for a wealthy economy that 'must continue to be reasonably equalitarian because its prosperity depends on mass purchasing power'. 1955-10-14 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Edwin G. Nourse · Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings Nourse, “Statement (former Chairman, Council of Economic Advisers), in Automation and Technological Change: Hearings” (1955-10-28) bears on M06b. Speaker: Edwin G. Nourse, former Chairman, Council of Economic Advisers (economist). Nourse rejects the view that technological progress 'takes care of its own economic problems'. Responding to members' and staff questions on automation-specific legislation. Missing: (M06b-02) A general call for generous policies to assimilate technological change, not a specific compensating policy. (M06b-04) Answers whether to legislate on 'the problems created by so-called automation'. He rejects legislation that would 'curb or redirect' discovery and its application, not compensating policy as such. Conditions: (M06b-02) Argued against the 'comfortable formula' that falling prices automatically restore jobs. That formula presupposes perfect competition and labour mobility, which do not hold in an economy of large corporations, unions and administered prices. (M06b-04) He immediately adds that every money bill, tax revision and regulatory measure gives 'punchcard or tape instructions' to the economy, and that general public policy should be framed accordingly. 1955-10-28 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 National Commission on Technology, Automation, and Economic Progress · Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I National Commission on Technology, Automation, and Economic Progress, “Technology and the American Economy: Report of the National Commission on Technology, Automation, and Economic Progress, Volume I” (1966-02) bears on M06b. The Commission recommends going beyond reform of existing social insurance to 'wholly new approaches to the problem of income maintenance'. Missing: Here the instrument is a response to poverty and the defects of means-tested assistance; this paragraph does not tie it to automation. Conditions: 'if found feasible', to be approached 'by stages'. The report names Milton Friedman, Robert Lampman and James Tobin as authors of concrete proposals. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M06b. Speaker: Wassily Leontief and Faye Duchin (economists). The authors call their study 'a first systematic albeit partial glance' at employment prospects nearly twenty years ahead. Missing: Names income distribution as a next research step (citing Duchin 1984); proposes no redistribution or compensating policy. Conditions: Part of a list of 'next steps in our agenda', with a closed dynamic model determining consumption and employment together. 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Gill A. Pratt · Is a Cambrian Explosion Coming for Robotics? Pratt, “Is a Cambrian Explosion Coming for Robotics?” (2015) bears on M06b. Pratt, a robotics researcher, reviews cloud robotics and deep learning and announces suggestions for addressing economic difficulties from robotics. Missing: The abstract does not say what the 'ways to address potential difficulties' are or tie them to distributional effects. 2015 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) bears on M06b. Attributes part of the productivity slowdown to automation pushed beyond the social optimum by the tax treatment of capital and labour-market rents. Missing: Identifies a tax bias as a cause of excessive automation and slow productivity growth; it is not argued here as compensation for distributional effects. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Automation and New Tasks: How Technology Displaces and Reinstates Labor Acemoglu and Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor” (2019-03) bears on M06b. The authors suggest policies that remove tax preferences for capital and rebalance technology towards new tasks. Missing: Proposes redirecting technology, not redistribution or compensation. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M06b. The authors identify tax treatment favouring capital as a distortion pushing technology toward automation. Missing: Concerns tax incentives for automation, not compensating policy for distributional effects. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on M06b. Cowen accepts a reason for larger cash transfers in AI-displacement scenarios but warns, citing resource-rich monarchies, that mass redistribution may create political problems. Missing: Weighs, rather than argues for, a compensating policy. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Betsey Stevenson · Artificial Intelligence, Income, Employment, and Meaning Stevenson, “Artificial Intelligence, Income, Employment, and Meaning” (2019-05) bears on M06b. Stevenson argues the gains from AI could compensate losers, but whether they are redistributed depends on politics; she notes current failure to redistribute technological gains. Missing: Does not argue for a specific redistributive or compensating policy. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyna Eloundou; Sam Manning; Pamela Mishkin; Daniel Rock · GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models Eloundou et al., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (2023-03-17) bears on M06b. The authors call for policy preparedness and point to prior work on education, training and safety-net reform, but decline to recommend specific policies. Missing: No specific redistributive or compensating policy is argued for. 2023-03-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on M06b. Acemoglu argues that new-task creation with better wage outcomes may need new institutions and regulation and a reorientation of the AI industry. Missing: Calls for policy to redirect technology rather than redistribution or compensation. 2024-05 Inspect claim & evidence ↗ Document page → Headlines replaced The report of the National Commission on Technology, Automation, and Economic Progress (1966), headlined as the earliest located tier-2 text. Earlier texts found by the re-checks of 2026-10-25 in the 1955 congressional hearings: the Diebold condensation of 14 October, on management’s share of responsibility for retraining and restoring the earning power of displaced workers, which counts as another compensating policy as the criterion is written, and Reuther’s statement of 17 October. Recorded 2026-10-25. M07 Exposure to automation measured from task contentCriterion. A text that estimates the share of jobs, occupations or tasks exposed or susceptible to computerisation or AI from task or occupation descriptions.
Earliest located, tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? · 2013-09-17 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (3) Articulates · Earliest located · Tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” (2013-09-17) articulates M07. The authors estimate occupations' susceptibility to computerisation by matching engineering bottleneck variables in O*NET task descriptions with a classifier trained on hand-labelled occupations. The headline estimate: 47 percent of US employment falls in occupations with high probability of computerisation. 2013-09-17 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Mika Pajarinen; Petri Rouvinen · Computerization Threatens One Third of Finnish Employment Pajarinen and Rouvinen, “Computerization Threatens One Third of Finnish Employment” (2014-01-13) articulates M07. Pajarinen and Rouvinen apply an occupation-level susceptibility estimate to Finnish employment and compare it with the United States. 2014-01-13 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Tyna Eloundou; Sam Manning; Pamela Mishkin; Daniel Rock · GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models Eloundou et al., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (2023-03-17) articulates M07. The authors estimate the share of US workers whose occupational tasks are exposed to LLMs, from O*NET task descriptions rated by humans and GPT-4. The authors describe their method: rating task exposure using a rubric applied by human annotators and GPT-4. Conditions: Exposure measure β (direct plus half of tool-dependent exposure); US occupations; GPT-4-assisted labels. 2023-03-17 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (4) Bears on · Tier 1 Mika Pajarinen; Petri Rouvinen · Computerization Threatens One Third of Finnish Employment Pajarinen and Rouvinen, “Computerization Threatens One Third of Finnish Employment” (2014-01-13) bears on M07. The authors caution that susceptibility is not a forecast of job loss. Missing: States the limits of the exposure measure rather than a measurement. 2014-01-13 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Jason Furman · Should We Be Reassured If Automation in the Future Looks Like Automation in the Past? Furman, “Should We Be Reassured If Automation in the Future Looks Like Automation in the Past?” (2019-05) bears on M07. Furman reports the range of published estimates of jobs at risk of automation and emphasises the wage/skill gradient of exposure. Missing: Reports others' exposure estimates; makes no estimate of its own. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyna Eloundou; Sam Manning; Pamela Mishkin; Daniel Rock · GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models Eloundou et al., “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (2023-03-17) bears on M07. The authors caution that exposure measures technical feasibility, not realised productivity or automation. Missing: A caveat on the measure rather than an estimate. 2023-03-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on M07. Acemoglu converts Eloundou et al.'s task exposure measure into a wage-bill-weighted share of US labour tasks exposed to AI. Missing: Re-aggregates another text's task-based exposure estimates rather than measuring exposure from task descriptions itself. 2024-05 Inspect claim & evidence ↗ Document page → M08 AI’s productivity effect over the next decadeCriterion. A text that gives a stated estimate of AI's effect on aggregate productivity or GDP over a horizon of about ten years from the text's date. The figure and horizon are recorded; the proposition is neutral as to size.
Earliest located, tier 1 Daron Acemoglu · The Simple Macroeconomics of AI · 2024-04-05 · wording attested 2025-01-06 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (1) Articulates · Earliest located · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-04-05) articulates M08. Acemoglu applies a Hulten's-theorem task framework to exposure and cost-saving estimates to bound AI's ten-year TFP effect, against much larger forecasts (Goldman Sachs, McKinsey). Allowing for hard-to-learn tasks where AI gains are smaller lowers the ten-year TFP bound. Acemoglu converts the TFP bounds to GDP effects under two investment-response assumptions. The author-hosted version of 5 April 2024 gives no more than 0.71% over ten years, and less than 0.55% allowing for hard tasks; the NBER version of May 2024 gives 0.66% and 0.53%. Conditions: (April version) Task-level cost savings from existing exposure and productivity estimates; the second figure adds the author’s easy/hard-task assumptions. (M08-04) Gains from cost savings at task level via automation and task complementarity only; excludes new tasks, deepening of automation and science-transforming effects; US economy. (M08-05) As above, plus easy/hard task classification assumptions. (M08-06) Modest investment response (capital rising in proportion to TFP) versus an investment boom like earlier automation. 2024-04-05 · wording attested 2025-01-06 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (2) Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M08. Speaker: Wassily Leontief and Faye Duchin (economists). Headline finding of the report. Missing: A labour-saving estimate over 20 years (to 2000), not a productivity or GDP estimate over about ten years, and it concerns computer-based automation, not AI. Conditions: Model-based projection (dynamic input-output model, 89 sectors, 53 occupations, scenario S3). The abstract adds consumption growth of about 2% a year through the 1980s and 0.5-1.1% through the 1990s, assuming full use of the projected labour force. 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Erik Brynjolfsson; Daniel Rock; Chad Syverson · Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics Brynjolfsson, Rock, Syverson, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics” (2017-11) bears on M08. Illustrates that a single AI application could add about 0.17 percentage points a year to US labour productivity growth over a decade. Second worked example showing that individual AI applications could add measurably to aggregate labour productivity over about a decade. Missing: (M08-02) An illustrative calculation for one AI application (autonomous vehicles), not an estimate of AI's aggregate effect. (M08-03) An illustrative calculation for one application (call centres), not an aggregate estimate. Conditions: (M08-02) Assumes drivers needed for current workload fall from 3.5 million to 1.5 million out of 122 million private nonfarm employment, over a supposed 10-year transition; excludes complementary gains. (M08-03) Assumes AI handles 60–70% of calls and reduces call-centre workers (about 2.2 million in 2015) by 60%. 2017-11 Inspect claim & evidence ↗ Document page → M09 Recent entry of machine intelligence into mainstream economicsCriterion. Not a criterion for a single passage. Assessed from the dated record: the earliest located tier-1 texts, within the searched scope, whose subject is AI, automation by AI or machine intelligence; plus explicit self-descriptions in tier-1 texts of the field as new or neglected. Tier-2 and outside texts are context.
Earliest located, tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" · 2019-05 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (1) Articulates · Earliest located · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) articulates M09. The editors open the volume by stating that economics research on AI is scarce and that existing work extrapolates from earlier technologies such as factory robots. The editors close by calling the economics of AI a new area of study and offering the volume as a starting point. 2019-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (25) Bears on · Tier 2 Joint Committee on the Economic Report, Subcommittee on Economic Stabilization · Automation and Technological Change: Hearings (statements by non-economists) Subcommittee on Economic Stabilization of the Joint Committee on the Economic Report, “Automation and Technological Change: Hearings (statements by non-economists)” (1955-10) bears on M09. Speaker: John Diebold, president, John Diebold & Associates (management consultant; not an economist witness). Diebold lists basic questions automation poses for 'economic theory, the way we measure productivity', and for cost accounting. Missing: A non-economist's call for economists to rethink theory in light of automation, not a description of the state of economics. It concerns automation in general, not machine intelligence. Conditions: Follows his question whether it is still meaningful to measure productivity by direct labour as the direct-labour content of manufacturing falls. 1955-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Walter S. Buckingham Jr. · Statement (testifying as an economist), in Automation and Technological Change: Hearings Buckingham, “Statement (testifying as an economist), in Automation and Technological Change: Hearings” (1955-10-14) bears on M09. Speaker: Walter S. Buckingham Jr., associate professor (industrial management), Georgia Institute of Technology; testifying as an economist. Introducing his written statement, Buckingham says he is neither an engineer nor an athlete and speaks for himself as an economist. The chairman introduces him as associate professor of industrial management who had spoken at the CIO automation conference in April 1955. Missing: Self-identification only; it establishes that an economist was the subcommittee's opening analytical witness on automation in October 1955, not what economics as a field held. Conditions: He adds that he speaks 'for myself only'. 1955-10-14 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Howard R. Bowen (ed.); Garth L. Mangum (ed.) · Automation and Economic Progress Bowen and Mangum, “Automation and Economic Progress” (1966) bears on M09. Speaker: Howard R. Bowen and Garth L. Mangum, editors (Editors' Introduction). The editors explain the Commission's origins in the persistent unemployment after the Korean conflict. They define automation technically as electronic feedback control of mechanical processes, which 'in common parlance' became a byword for anything reducing employment. Missing: Describes the early-1960s policy debate on unemployment, not economics as a research field or machine intelligence. Conditions: Alongside the 'structural unemployment' and 'deficiency of aggregate demand' hypotheses. 1966 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Edwin Mansfield · Technological Change: Measurement, Determinants, and Diffusion Mansfield, “Technological Change: Measurement, Determinants, and Diffusion” (1966-02) bears on M09. Speaker: Edwin Mansfield, University of Pennsylvania (economist). Mansfield reviews output-per-man-hour evidence (Clague and Greenberg, Kendrick and Sato, CEA) for acceleration. Missing: Shows that economists were debating whether automation-era technical change was faster, but is not a statement about machine intelligence as a research topic. Conditions: 'and it is equally possible to find statements asserting the opposite'. He quotes Diebold and DuBridge (non-economists) for the opposite view. 1966-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 · Relay of another author Nils J. Nilsson · Artificial Intelligence, Employment, and Income Nilsson, “Artificial Intelligence, Employment, and Income” (1984) bears on M09. Nilsson records that he consulted economists (Leontief, Duchin, Simon, Solow) and that Solow doubted his premises; the body sets Simon's and Leontief's opposing views against each other. Missing: Reports one economist's reaction; not a self-description of the field as new or neglected. 1984 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Wassily Leontief; Faye Duchin · The Impacts of Automation on Employment, 1963–2000. Final Report Leontief and Duchin, “The Impacts of Automation on Employment, 1963–2000. Final Report” (1984-04) bears on M09. Speakers: Wassily Leontief and Faye Duchin (economists), Institute for Economic Analysis, New York University; Wassily Leontief and Faye Duchin (economists). Opening of the 1984 NSF-funded final report's abstract. Scope limitation in the report's summary chapter. Missing: (M09-06) Characterises the literature as divided, not new or neglected. The subject is computer-based automation, not machine intelligence as such. (M09-07) Names AI-type capabilities only to exclude them from the modelled scenarios. It is evidence that a 1984 economic study scoped machine intelligence out, not that it analysed it. Conditions: (M09-06) The authors argue that policy should rest on 'concrete, detailed analysis' rather than opinion. (M09-07) The authors say they 'explicitly excluded from our scenarios any major' … 'break-throughs in computer technology that might affect significant numbers of workers before the year 2000' (the sentence runs from p. 1.32 to p. 1.34 across the Figure 1.4 page). 1984-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Timothy Bresnahan; Erik Brynjolfsson; Lorin Hitt · Information Technology, Workplace Organization and the Demand for Skilled Labor: Firm-Level Evidence Bresnahan, Brynjolfsson, Hitt, “Information Technology, Workplace Organization and the Demand for Skilled Labor: Firm-Level Evidence” (1999-05) bears on M09. The authors summarise the economics literature of the 1990s: the unexplained shift in labour demand is attributed to skill-biased technical change, which many link to information technology. Missing: Describes economists' consensus on technology and wage inequality, not the state of research on machine intelligence. 1999-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Jeffrey D. Sachs; Laurence J. Kotlikoff · Smart Machines and Long-Term Misery Sachs and Kotlikoff, “Smart Machines and Long-Term Misery” (2012-12) bears on M09. The authors characterise their profession as reluctant to treat machines as competitors for labour; elsewhere (pdf p. 7) they attribute this to the record of real wage growth and reliance on Cobb-Douglas production functions. Missing: Concerns machines and automation in general, not machine intelligence as a research topic; no claim about when the field engaged it. 2012-12 · wording attested 2013-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Carl Benedikt Frey; Michael A. Osborne · The Future of Employment: How Susceptible Are Jobs to Computerisation? Frey and Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” (2013-09-17) bears on M09. The authors present their estimate as filling a gap in the literature on recent technological progress and employment. Missing: A gap claim about one kind of quantification, not a description of the field as new. 2013-09-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Joel Mokyr; Chris Vickers; Nicolas L. Ziebarth · The History of Technological Anxiety and the Future of Economic Growth: Is This Time Different? Mokyr, Vickers, Ziebarth, “The History of Technological Anxiety and the Future of Economic Growth: Is This Time Different?” (2015) bears on M09. Mokyr, Vickers and Ziebarth survey economists' historical views on technological unemployment, moral effects of technology and stagnation, and compare them with current concerns. Missing: Concerns economists' past views of technology and machinery in general, not machine intelligence; no self-description of the field as new or neglected. 2015 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Seth G. Benzell; Laurence J. Kotlikoff; Guillermo LaGarda; Jeffrey D. Sachs · Robots Are Us: Some Economics of Human Replacement Benzell et al., “Robots Are Us: Some Economics of Human Replacement” (2015-02) bears on M09. The authors describe a shift from the postwar dismissal of technological-unemployment concerns to recent economic work on smart machines, citing Brynjolfsson and McAfee and Aghion et al. as part of a 'burgeoning literature'. Missing: Describes a renewed debate with acknowledged older roots, not the field as new. 2015-02 · wording attested 2018-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on M09. Nordhaus notes the scarcity of macroeconomic work on accelerating growth from computation and AI, apart from endogenous-growth curiosities and Simon's earlier analysis. Introducing computer scientists' views of AI and superintelligence, Nordhaus signals that these ideas lie outside economists' usual frame. Nordhaus credits Simon's 1965 book with a macroeconomic analysis of computerisation based on the factor-price frontier, and criticises its assumption of a constant capital share. Missing: (M09-11) The claim is about singularity in modern macroeconomics, after a substantive account of Simon; it cannot establish recent entry of machine intelligence generally. (M09-12) Describes economists' unfamiliarity with the material, not explicitly the field as new or neglected. (M09-35) A 2015 economist's account of earlier economic analysis of computerisation, not a description of the field as new; it bears on M09 as counter-context to an "only recently" reading. 2015-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment Acemoglu and Restrepo, “The Race Between Machine and Man: Implications of Technology for Growth, Factor Shares and Employment” (2016-05) bears on M09. After citing concerns that robotics and AI will leave workers unable to compete, the authors say no comprehensive framework for these effects exists. Missing: A single self-description of a gap in theory, not a dated record of the field; it does not say economists ignored the topic. 2016-05 · wording attested 2017-06 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones, Jones, “Artificial Intelligence and Economic Growth” (2017-10) bears on M09. Presents the paper as agenda-setting for research on AI and growth. Missing: Implies the agenda is still to be shaped but does not state that the topic was neglected. 2017-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Erik Brynjolfsson; Daniel Rock; Chad Syverson · Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics Brynjolfsson, Rock, Syverson, “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics” (2017-11) bears on M09. Contrasts technologist optimism with the productivity pessimism the authors attribute to economists and statisticians, and argues both can be right. Missing: Characterises economists' outlook on technology and growth, not the field's engagement with machine intelligence as a topic. 2017-11 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) bears on M09. Argues economists must say how gains from AI can be shared, not merely that they could be. Missing: Criticises a habitual stance of economists rather than describing the field's engagement with AI as recent or neglected. 2017-12 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · Artificial Intelligence, Automation and Work Acemoglu and Restrepo, “Artificial Intelligence, Automation and Work” (2018-01) bears on M09. Opens the conclusion by saying the debate on automation and the future of work lacks a satisfactory framework, which the authors blame for a false dichotomy. The authors describe two sides of the debate, alarmist predictions of the end of work and economists' reassurance from history, and reject both as a false dichotomy. Missing: (M09-17) Self-description of a theoretical gap about automation and work; not a dated claim about when economists took up AI. (M09-18) Describes a view the authors attribute to many economists (and reject as one side of a false dichotomy); it is not the authors' own view and not a dated record. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Manuel Trajtenberg · AI as the Next GPT: A Political-Economy Perspective Trajtenberg, “AI as the Next GPT: A Political-Economy Perspective” (2018-01) bears on M09. Self-critique of economists' tendency to treat GDP growth as welfare and give only lip service to distribution. Says economics has traditionally neglected the direction of technical change, motivating the human-enhancing versus human-replacing distinction. Missing: (M09-19) Describes economists' neglect of losers from technical change, not of machine intelligence as a topic. (M09-20) Concerns the economics of technical change in general, not machine intelligence. 2018-01 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb · Economic Policy for Artificial Intelligence Agrawal, Gans, Goldfarb, “Economic Policy for Artificial Intelligence” (2018-06) bears on M09. Summarises economists' long-run optimism on jobs while stressing short-run adjustment costs. Missing: Characterises economists' views on AI and jobs; says nothing about when the field took up AI. 2018-06 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Jason Furman; Robert Seamans · AI and the Economy Furman and Seamans, “AI and the Economy” (2018-06) bears on M09. Concludes that economic research on AI's productivity and labour effects has only begun and calls for more empirical work. Explains why the NBER volume's chapters rely on theoretical models and national statistics rather than firm-level data. Missing: (M09-21) A self-description of the research as early; not a dated record. (M09-22) Describes a data gap explaining the literature's reliance on theory and aggregate statistics; not a statement about when the field began. 2018-06 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu; Pascual Restrepo · The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand Acemoglu and Restrepo, “The Wrong Kind of AI? Artificial Intelligence and the Future of Labor Demand” (2019-03) bears on M09. The authors describe how economists studying AI framed it in 2019. The authors note that economics (Herbert Simon) was among the disciplines that pioneered AI in the 1950s. Missing: (M09-25) Says nothing about when economists began studying AI. (M09-26) Indicates early economist participation in AI research, which complicates but does not test the recency claim. 2019-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans, Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) bears on M09. The editors state that the NBER convened its first conference on the economics of AI in September 2017 to set a research agenda. The editors explain that the volume's growth and labour chapters are speculative because AI had not yet diffused, forcing reliance on theory or proxy technologies. Missing: (M09-28) Dates an institutional event rather than describing the field as new; supports M09 only as one dated marker. (M09-29) Describes the evidence base as thin because of limited diffusion, not the field as recently engaged. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on M09. Cowen says economists have done little work on what redistribution must mean in AI-transformed futures. Missing: Neglect claim is confined to redistribution beyond the Paretian model, not economics of AI generally. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Betsey Stevenson · Artificial Intelligence, Income, Employment, and Meaning Stevenson, “Artificial Intelligence, Income, Employment, and Meaning” (2019-05) bears on M09. Stevenson describes economists' prevailing view, grounded in historical experience and IGM panel responses, that technological change has not reduced employment. Missing: Describes economists' earlier views on technology and employment, not the recency of engagement with AI. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Emilio Calvano; Giacomo Calzolari; Vincenzo Denicolò; Sergio Pastorello · Artificial Intelligence, Algorithmic Pricing, and Collusion Calvano et al., “Artificial Intelligence, Algorithmic Pricing, and Collusion” (2020-10) bears on M09. The abstract shows a 2020 AER article whose subject is AI pricing algorithms in oligopoly. Missing: A dated data point, not a statement about the field's recency. 2020-10 Inspect claim & evidence ↗ Document page → Headlines replaced Nordhaus (2015), headlined as earliest located. Reclassified as bearing on it after the cross-family check: his remark concerns the singularity in macroeconomics. The introduction to the NBER volume of 2019 is now earliest located. Recorded 2026-10-12. E · Superintelligence and existential risk Whether economists took up machine intelligence that improves itself, machine intelligence beyond human level in essentially all domains and existential risk from AI, how developers’ incentives bear on safety, and what texts say about the state of economics on these questions. Texts from 1998 to 2026, mostly NBER working papers, with Robin Hanson’s texts of 1998 to 2000 and posts from his 2008 debate with Eliezer Yudkowsky. E05 counts only explicit statements about the state of economics; economists’ engagement itself is recorded under E01 to E04.
E01 Machine intelligence improving itself accelerates its capability growthCriterion. A text that states or models AI systems improving AI (their own capabilities, their successors' design, or AI research itself) so that capability growth accelerates. Accelerating output growth from automated idea production without AI improving AI is G02 and only bears on E01.
Earliest located, tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth · 2017-10 ↗
Earliest located, tier 2 Eliezer Yudkowsky · Recursive Self-Improvement · 2008-12-01 ↗
Earliest located tier-1 text that contests it Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth · 2017-10 ↗
Earliest located tier-2 text that contests it Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") · 1998-10 · wording attested 2000-11-17 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (9) Contests · Earliest located in tier 2 · Tier 2 · Engages a premise Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") Hanson, “Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept")” (1998-10) contests E01. Hanson's main objection to an intelligence explosion: diminishing returns to self-improvement, illustrated by schooling and factory learning curves. 1998-10 · wording attested 2000-11-17 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 · Relay of another author Robin Hanson · AI Go Foom Hanson, “AI Go Foom” (2008-11-20), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E01. Hanson summarises Yudkowsky's argument "as best I can, so I can then respond". 2008-11-20 Inspect claim & evidence ↗ Later texts by the same author that contest it Robin Hanson · Wrapping Up · 2008-12-07 ↗ Document page → Articulates · Earliest located in tier 2 · Tier 2 Eliezer Yudkowsky · Recursive Self-Improvement Yudkowsky, “Recursive Self-Improvement” (2008-12-01), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E01. Yudkowsky's statement of the claim Hanson contests; the post notes that Hanson "objected to the 'local' part". 2008-12-01 Inspect claim & evidence ↗ Document page → Contests · Tier 2 · Engages a premise Robin Hanson · Wrapping Up Hanson, “Wrapping Up” (2008-12-07), in “The Hanson-Yudkowsky AI-Foom Debate”, contests E01. Hanson's objection to Yudkowsky's "growth math"; he repeats it in "What Core Argument?" (10 December 2008). 2008-12-07 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Relay of another author William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates E01. Nordhaus quotes I. J. Good to explain the Singularity idea. 2015-09 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 · Model scenario Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones and Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates E01. The authors model "a unit of A.I. working to improve itself" by a differential equation for cognitive productivity. The authors cite Good (1965), Vinge (1993) and Kurzweil (2005) and point to Nordhaus (2015) for an economic discussion. Conditions: Self-improvement law dA/dt = A^(1+ω) for cognitive productivity, with ω > 0. 2017-10 Inspect claim & evidence ↗ Document page → Contests · Earliest located · Tier 1 · Engages a premise Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones and Jones, “Artificial Intelligence and Economic Growth” (2017-10) contests E01. The authors list bottlenecks that could prevent a singularity even with complete automation. 2017-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) articulates E01. Discussing AI safety, the authors say models in which capital recursively self-improves without human input are the most worrying for existential risk. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) articulates E01. Abstract sentence contrasting automation of production (self-replication) with automation of R&D (self-improvement). Citing the high capital share in AI R&D (Besiroglu et al., 2024), the authors describe a feedback in which automating AI R&D speeds AI capability growth, then cite calibrated models of this feedback (Erdil et al. 2025; Lifland et al. 2025) as having 'speculative elements'. In the risk subsection, the authors contrast a scenario of acceleration from R&D automation alone with scenarios of automated physical production, each with distinct safety concerns. Conditions: (E01-28) Within the reviewed growth models; acceleration not sufficient 'in isolation'. (E01-29) 'could'; hedged 'in some sense'. (E01-31) Scenario; 'potential for'. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (14) Bears on · Tier 1 Robin Hanson · Is a Singularity Just Around the Corner? What it takes to get explosive economic growth Hanson, “Is a Singularity Just Around the Corner? What it takes to get explosive economic growth” (1998-06) bears on E01. Hanson opens his journal article by reporting the singularity as others' speculation and recasting it as a question about growth rates. Hanson's bottleneck argument against a technology-driven sudden growth explosion. Missing: (E01-01) No machine intelligence improving itself or its successors; "singularity" is framed purely as a growth-rate episode. (E01-02) Addresses a growth explosion from any single technology; no AI improving AI. 1998-06 · wording attested 1999-01-28 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept") Hanson, “Some Skepticism (comment in "A Critical Discussion of Vinge's Singularity Concept")” (1998-10) bears on E01. Hanson tests a growth-rate reading of Vinge against growth economics: a one-time intelligence gain gives a one-time jump in output, not faster growth. Missing: Disputes a link between intelligence and growth rates, not AI improving AI. 1998-10 · wording attested 2000-11-17 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) bears on E01. In an endogenous learning-by-doing variant, the model loses steady-state solutions and growth keeps accelerating. Missing: Accelerating growth via endogenous technology; no machine intelligence improving its own design. Conditions: Learning-by-doing technology with parameters such that the expression for steady growth is not positive. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Setting the Stage Hanson, “Setting the Stage” (2008-11-18), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E01. Hanson summarises points of agreement and disagreement with Yudkowsky at the start of the debate. Missing: Puts a low probability on one hand-coded system going suddenly from nearly powerless to overwhelmingly powerful; it does not deny that AI improving AI could accelerate capability growth. 2008-11-18 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Debating Yudkowsky Hanson, “Debating Yudkowsky” (2011-07-03), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E01. Hanson concludes his five reasons why powerful architectural insights are rare. Missing: Doubts that one team's architectural innovation could take it from tiny to world takeover within weeks; the passage accepts future superintelligences and does not deny accelerating AI self-improvement in general. 2011-07-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Foom Debate, Again Hanson, “Foom Debate, Again” (2013-02-18), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E01. Hanson restates his position in 2013, granting the novelty of self-modification while doubting that it produces a foom. Missing: A concession that intentional mind redesign could matter, followed at once by "seems unlikely to generate foom". 2013-02-18 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on E01. Nordhaus summarises his tests: five of seven are negative, and the two positive ones put the Singularity 100 years or more away. Missing: A growth-singularity test (growth above 20% a year), not AI improving AI (scope boundary). Conditions: Simple extrapolation of the two positive tests. 2015-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Ari Hyytinen · Ovatko hyvät johtamismenetelmät talouskasvun lähde? (pääkirjoitus) Hyytinen, “Ovatko hyvät johtamismenetelmät talouskasvun lähde? (pääkirjoitus)” (2017) bears on E01. Hyytinen defines the "economic-growth singularity". Missing: AI accelerates general technological progress; no AI improving AI. 2017 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Joshua S. Gans · Self-Regulating Artificial General Intelligence Gans, “Self-Regulating Artificial General Intelligence” (2018-02) bears on E01. After quoting Good and citing Yudkowsky (2007) and Omohundro (2007) on "recursive self-improvement", Gans recasts it in economic terms. Gans's main result: a paperclip AI refrains from activating power-accumulating successor AIs. Missing: (E01-20) States the mechanism of self-improvement (building a better AI) without acceleration; the paper argues that a control problem limits it. (E01-21) Limits the creation of power-accumulating successors; the model still has the AI activate research-specialist AIs, so accelerating cognitive self-improvement is not contested. Conditions: Assumptions A1 and A2. 2018-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) bears on E01. The abstract distinguishes AI letting capital self-replicate from AI letting it self-improve via knowledge production. Missing: Self-improvement of capital through knowledge production, framed as growth, not AI improving AI's own capabilities. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) bears on E01. Opening paragraph: the authors say future models may reach AGI and, unlike earlier technologies, could automate R&D itself. The clause names AI creating better AI systems. The authors characterise the 'singularity' hypothesis and add that absent current economic effects cannot refute it because the theory predicts little advance warning. Second of the review's 'robust findings': automating R&D in isolation has ambiguous growth effects. Missing: (E01-27) AI building better AI, without stated acceleration of capabilities. (E01-30) Describes a hypothesis held by others and what it predicts; it does not itself state that AI improves AI, and the authors neither endorse nor reject the hypothesis here. (E01-32) About output growth from automated research in general (G02 territory), not about AI improving AI; it bounds the acceleration rather than rejecting it. Conditions: (E01-27) Conditional on future models reaching AGI ('may reach'). (E01-32) Automating R&D in isolation; depends on parallelizability and research constraints. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Charles I. Jones · The A.I. Dilemma: Growth versus Existential Risk Jones, “The A.I. Dilemma: Growth versus Existential Risk” (2023-11) bears on E01. Cites Aghion, Jones and Jones (2019), Trammell and Korinek (2020), Davidson (2021), Nordhaus (2021), Erdil and Besiroglu (2023). Missing: Growth explosion from automated idea production (G02); does not state that AI improves AI. Conditions: 'Once machines can produce ideas'; 'could', 'potentially'. 2023-11 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Anton Korinek; Donghyun Suh · Scenarios for the Transition to AGI Korinek and Suh, “Scenarios for the Transition to AGI” (2024-03) bears on E01. Following Aghion et al. (2019) and Trammell and Korinek (2023), the model shows a 'type II singularity' (infinite output in finite time) once automation passes a threshold. Missing: Self-reinforcing growth from automated R&D (G02); AI does not improve AI in the model. Conditions: Sufficient automation of final-output and knowledge production; constant saving rate and allocation; no depreciation. 2024-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on E01. Explains the paper's scope; the example given is DeepMind's protein-folding and crystal-structure work. Missing: About AI changing scientific research generally, not AI improving AI; a timing judgement for the paper's horizon, not a rejection of the possibility. Conditions: Within a 10-year horizon. 2024-05 Inspect claim & evidence ↗ Document page → E02 Machine intelligence exceeding human cognition in essentially all domainsCriterion. A text that states or models the possibility of machine intelligence that surpasses humans across essentially all cognitive tasks (superintelligence, AGI beyond human level). Machines doing all economically valuable work more cheaply is M05 and bears on E02 only if cognitive superiority is stated.
Earliest located, tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) · 2023-10 · wording attested 2023-11-29 ↗
Earliest located, tier 2 Robin Hanson · Debating Yudkowsky · 2011-07-03 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (6) Articulates · Earliest located in tier 2 · Tier 2 Robin Hanson · Debating Yudkowsky Hanson, “Debating Yudkowsky” (2011-07-03), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E02. Hanson's summary position after the 2011 debate. 2011-07-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Relay of another author William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates E02. Nordhaus reports the computer-science view, quoting Bostrom (2006). 2015-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Relay of another author Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates E02. Korinek and Stiglitz open their speculative long-run section with expert forecasts (citing Bostrom 2014). Conditions: If human-level AGI is reached. 2017-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Relay of another author Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans and Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) articulates E02. The editors distinguish machine learning, AGI and automation as meanings of AI. 2019-05 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) articulates E02. Explaining their focus on average wages and the labour share, the authors state that the likeliest transformative long-run scenario is AI outsmarting all humans. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Relay of another author Charles I. Jones · The A.I. Dilemma: Growth versus Existential Risk Jones, “The A.I. Dilemma: Growth versus Existential Risk” (2023-11) articulates E02. Preceded by 'many experts worry that these advances entail existential risk:'. Jones reports warnings from 'a substantial contingent of the A.I. research community, including leading researchers at OpenAI and Google'. 2023-11 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (9) Bears on · Tier 1 · Relay of another author Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) bears on E02. Hanson motivates his model of machine intelligence by the stated goals of AI and robotics. Missing: All human cognitive tasks, but no claim that machines surpass humans (M05 territory). 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones and Jones, “Artificial Intelligence and Economic Growth” (2017-10) bears on E02. The authors note that their "labor free" singularities need complete automation of cognitive tasks. Missing: Human-level, not beyond-human, intelligence. 2017-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) bears on E02. Definition of AGI used in the review's introduction. Missing: Defines AGI as parity with skilled humans on all cognitive tasks ('at least as well as'); it does not state that machines would exceed human cognition. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Charles I. Jones · The A.I. Dilemma: Growth versus Existential Risk Jones, “The A.I. Dilemma: Growth versus Existential Risk” (2023-11) bears on E02. Jones's own statement of possibility, followed by growth models in which machines produce ideas. Missing: 'Many cognitive tasks', not essentially all; 'innovating itself' means AI producing ideas, not explicitly improving AI. 2023-11 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Anton Korinek; Donghyun Suh · Scenarios for the Transition to AGI Korinek and Suh, “Scenarios for the Transition to AGI” (2024-03) bears on E02. The paper models wages and output under scenarios that 'may culminate' in AGI. Missing: Defines AGI as ability to perform all human tasks (M05-type), not cognitive superiority. 2024-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on E02. Footnote 1 cites Korinek and Suh (2024) and those who see a confirmation of Kurzweil's (2005) singularity forecast; Acemoglu contrasts 'more grounded' forecasters. Missing: Reports others' AGI expectations (all tasks, M05-type); no cognitive superiority stated and not Acemoglu's view. 2024-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Charles I. Jones · How Much Should We Spend to Reduce A.I.'s Existential Risk? Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) bears on E02. The preceding sentence reports experts on catastrophic outcomes 'either from malicious use or even from a superintelligent A.I. itself'. Missing: 'Most tasks', not essentially all; framed as world-changing benefit. 2025-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Ajay Agrawal; Joshua Gans; Avi Goldfarb · Genius on Demand: The Value of Transformative Artificial Intelligence Agrawal, Gans and Goldfarb, “Genius on Demand: The Value of Transformative Artificial Intelligence” (2025-10) bears on E02. Model of 'genius on demand' AI; human geniuses specialise where their comparative advantage over AI is greatest. Missing: AI approaching, not exceeding, human genius efficiency; in knowledge work only. Conditions: 'if AI efficiency approaches human genius efficiency'. 2025-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Pascual Restrepo · We Won't be Missed: Work and Growth in the AGI World Restrepo, “We Won't be Missed: Work and Growth in the AGI World” (2025-10) bears on E02. Restrepo's chapter for the NBER volume The Economics of Transformative AI; the abstract derives labour-share convergence to zero. Missing: All economically valuable work done by compute (M05); no statement that machine cognition surpasses human cognition. 2025-10 Inspect claim & evidence ↗ Document page → E03 Existential risk from AI as a matter for economic analysisCriterion. A text that treats existential or catastrophic risk from AI (extinction, loss of control, misaligned goals) as a quantity or consideration in economic analysis, welfare trade-offs or policy.
Earliest located, tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth · 2015-09 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (8) Articulates · Earliest located · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates E03. Nordhaus adds the superintelligent machines themselves to his list of risks in a section on "evil agents". Conditions: If Good's description of superintelligence is taken seriously. 2015-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates E03. The authors discuss whether humans can keep property rights over AI entities. Drawing on Korinek (2017), the authors describe a Malthusian race between humans and autonomous AI entities. Conditions: (E03-05) If the intelligence gap between humans and super-intelligent AI entities grows too large. (E03-06) Humans do not or cannot assert ownership over AI entities; AI reproduction technology improves exponentially. 2017-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Joshua S. Gans · Self-Regulating Artificial General Intelligence Gans, “Self-Regulating Artificial General Intelligence” (2018-02) articulates E03. Gans takes Bostrom's control problem and "paperclip apocalypse" into economic theory. Gans states the scope of his self-regulation result. Conditions: Assumptions A1 (an AI self-improves only by employing AIs with targeted goals) and A2 (improving power requires power-accumulation capabilities). 2018-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) articulates E03. The authors frame AI safety as a public good to which public-goods and climate-risk economics could apply. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) articulates E03. Concluding subsection placing AI existential risk within the paper's framing of economic transformation; it goes on to say the scenarios carry different risk profiles. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Charles I. Jones · The A.I. Dilemma: Growth versus Existential Risk Jones, “The A.I. Dilemma: Growth versus Existential Risk” (2023-11) articulates E03. The abstract sets growth gains against existential risk, which it attributes to 'many experts', and frames the paper as a welfare analysis of whether to continue or stop. First of four insights: 'With log utility, the models are remarkably unconcerned with existential risk'. Insight 2 says CRRA utility with γ of 2 or more is 'quite conservative'. Key Point 3: with γ ≤ 1 any risk short of sure annihilation is acceptable for an immediate singularity; with γ = 2 or 3 the cutoffs are much smaller. The 'key exception' to the conservative result: mortality gains are 'in the same units' as existential risk. Conditions: (E03-11) Log utility; simple static model. (E03-12) γ = 3; immediate singularity with infinite consumption; no A.I. mortality improvement; ρ − b = 1%, g0 = 2%, m0 = 1%, v(c0) = 6. (E03-13) If A.I.-driven innovation extends life expectancy and reduces mortality. 2023-11 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Charles I. Jones · The AI Dilemma: Growth versus Existential Risk Jones, “The AI Dilemma: Growth versus Existential Risk” (2024-12) articulates E03. Published version of Jones's working paper; the abstract attributes existential risk from a misaligned superintelligence to 'many experts'. 2024-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Charles I. Jones · How Much Should We Spend to Reduce A.I.'s Existential Risk? Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E03. Sentence continues 'in the Covid-19 pandemic'; Jones adds that mortality risk to people alive 'more than justifies spending large amounts'. Abstract follows a back-of-envelope calculation from a $10 million value of life and a 1% mortality risk. Preceded by the note that 12–33% of runs lead to zero mitigation investment (very small or hard-to-mitigate risk). After saying that valuing future generations would raise the amount. Conditions: (E03-18) 'For most of the scenarios and parameter combinations considered'; no weight on future generations. (E03-19) Monte Carlo average over assumed parameter distributions; zero altruism to future generations. 2025-03 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (5) Bears on · Tier 2 Robin Hanson · Setting the Stage Hanson, “Setting the Stage” (2008-11-18), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E03. First item in Hanson's list of points on which he and Yudkowsky "seem to agree". Missing: Risk is not specified as extinction or loss of control and is not placed in an economic analysis. 2008-11-18 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Wrapping Up Hanson, “Wrapping Up” (2008-12-07), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E03. Hanson argues that property-rights institutions protect humans in his preferred emulation scenario. Missing: Concerns emulations rather than misaligned AI; it argues that extermination is unlikely and puts no risk into a welfare or policy analysis. 2008-12-07 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) bears on E03. The footnote qualifies the authors' assumption that "our social and economic system will be maintained" after AGI. Missing: Reports AI-safety researchers' doom scenarios in a footnote; the main text assumes them away. 2017-12 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Relay of another author Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans and Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) bears on E03. The editors summarise Patrick Francois's comment on a world where machines are better at all tasks. Missing: Value specification as a political task; no extinction or loss-of-control risk is stated. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Joshua S. Gans · Can Socially-Minded Governance Control the AGI Beast? Gans, “Can Socially-Minded Governance Control the AGI Beast?” (2023-11) bears on E03. Refers to the 'potential existential risks' that motivated OpenAI's funders; the model then assumes the risks are real. Missing: A literature and motivation statement, not a substantive economic treatment of the risk; the model treats harm from dangerous users as an expected negative externality (E > 0), not as extinction or loss of control. 2023-11 Inspect claim & evidence ↗ Document page → E04 AI developers’ incentives matter for safetyCriterion. A text that analyses how incentives, competition or governance of AI developers affect the safety or control of advanced AI.
Earliest located, tier 1 Joshua S. Gans · Can Socially-Minded Governance Control the AGI Beast? · 2023-11 ↗
Earliest located, tier 2 Robin Hanson · Friendliness Factors · 2008-11-16 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (6) Articulates · Earliest located in tier 2 · Tier 2 Robin Hanson · Friendliness Factors Hanson, “Friendliness Factors” (2008-11-16), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E04. Hanson lists competition factors (info leaks, shared standards, anti-trust, sharing deals) that affect how far a winner leads. 2008-11-16 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Robin Hanson · Friendly Projects vs. Products Hanson, “Friendly Projects vs. Products” (2008-12-05), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E04. Hanson separates "project friendliness" (how race winners treat losers) from "product friendliness" (whether creations rebel). Hanson qualifies his claim that developers have strong incentives to avoid rebellious creations. 2008-12-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 2 Robin Hanson · Wrapping Up Hanson, “Wrapping Up” (2008-12-07), in “The Hanson-Yudkowsky AI-Foom Debate”, articulates E04. Hanson proposes liability and insurance for AI developers as the institutional answer to the friendliness problem. 2008-12-07 Inspect claim & evidence ↗ Document page → Articulates · Earliest located · Tier 1 · Model scenario Joshua S. Gans · Can Socially-Minded Governance Control the AGI Beast? Gans, “Can Socially-Minded Governance Control the AGI Beast?” (2023-11) articulates E04. Abstract: a model under idealised conditions (AGI risks real, safe AGI possible, socially-minded funders exist) motivated by OpenAI's governance. Case in which the for-profit entity is viable in direct competition; 'they' is the socially-minded entity. Conclusion; Gans adds that whether national or global intervention is needed is left to future research, while the results point to the latter. Conditions: (E04-10) Idealised assumptions; non-excludable technology; financial-viability constraint. (E04-11) For-profit entity viable in competition with the socially-minded entity. 2023-11 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Charles I. Jones · How Much Should We Spend to Reduce A.I.'s Existential Risk? Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E04. Listed among 'many related questions that merit further study'. 2025-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 · Model scenario Ethan Bueno de Mesquita; Wioletta Dziuda; Mattias Polborn · The AGI Race and Existential Risk Bueno de Mesquita, Dziuda and Polborn, “The AGI Race and Existential Risk” (2026-05) articulates E04. Model of firms allocating scarce resources between speed and safety in a race to AGI; the abstract opens by questioning the common assumption that competition and resources increase risk. Holds above a critical market size; below it firms have positive expected payoff from achieving AGI. The model's risk variable is 'doom risk'. Conditions: Market size above the model's critical cutoff. 2026-05 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (6) Bears on · Tier 2 Robin Hanson · Emulations Go Foom Hanson, “Emulations Go Foom” (2008-11-22), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E04. Hanson's scenario of competing whole-brain-emulation ("bot") projects. Missing: The race is analysed for leads and dominance; the safety or control of the systems is not discussed here. 2008-11-22 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Wrapping Up Hanson, “Wrapping Up” (2008-12-07), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E04. Hanson on why Yudkowsky's concern feels urgent. Missing: A counterfactual about regulation, not an analysis of developers' incentives. 2008-12-07 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) bears on E04. Nordhaus compares the diffusion of superintelligence technology with that of weapons technology. Missing: Diffusion to adversaries, not how developers' incentives or governance affect safety. 2015-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) bears on E04. The authors consider halting AI progress before turning to a human "reservation" policy. Missing: Asks whether humanity should slow or halt AI progress; does not analyse developers' incentives. 2017-12 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 · Model scenario Joshua S. Gans · Self-Regulating Artificial General Intelligence Gans, “Self-Regulating Artificial General Intelligence” (2018-02) bears on E04. Gans draws a governance implication from his model. Missing: Governance of AI property, not developers' incentives or competition. 2018-02 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) bears on E04. Closing sentence of the dangerous-AI subsection. Missing: Lists safety research and governance frameworks as policy responses; it does not analyse developers' incentives, competition or governance structures. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → E05 Economics’ engagement with superintelligence, AGI and the singularityCriterion. A passage articulates E05 only if it explicitly says something about the state of economics on the topic: that it is new, neglected, speculative, now on the agenda of economics, or beyond economists' existing tools. Economists' engagement itself is recorded under E01–E04 and in the timeline (the dated record), not here. Machine intelligence that does human jobs is not counted as AGI for E05, as for E02. Maintainer decision of 2026-10-15.
Earliest located, tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth · 2015-09 ↗ 1952 1966 1984 2000 2015 2024 Filled dots articulate, open dots contest, small dots bear on the proposition; a large dot is an earliest located text. A line below a dot runs to the date from which its wording is attested. Texts that articulate or contest it (9) Articulates · Earliest located · Tier 1 William D. Nordhaus · Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth Nordhaus, “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth” (2015-09) articulates E05. Nordhaus introduces the computer-science literature on AI and the Singularity to economists. Nordhaus surveys economic antecedents (Keynes, Simon and others) before his model. 2015-09 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philippe Aghion; Benjamin F. Jones; Charles I. Jones · Artificial Intelligence and Economic Growth Aghion, Jones and Jones, “Artificial Intelligence and Economic Growth” (2017-10) articulates E05. The abstract lists singularities and superintelligence among the questions for an economics research agenda. 2017-10 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Anton Korinek; Joseph E. Stiglitz · Artificial Intelligence and Its Implications for Income Distribution and Unemployment Korinek and Stiglitz, “Artificial Intelligence and Its Implications for Income Distribution and Unemployment” (2017-12) articulates E05. The authors justify taking up AGI in an economics paper. The authors flag the speculative nature of their AI-entity scenario. 2017-12 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Joshua S. Gans · Self-Regulating Artificial General Intelligence Gans, “Self-Regulating Artificial General Intelligence” (2018-02) articulates E05. Gans explains why he takes a different modelling route. 2018-02 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Ajay Agrawal; Joshua Gans; Avi Goldfarb · Introduction to "The Economics of Artificial Intelligence: An Agenda" Agrawal, Gans and Goldfarb, “Introduction to "The Economics of Artificial Intelligence: An Agenda"” (2019-05) articulates E05. The editors justify including superintelligence in an economics agenda. The editors characterise how much the volume covers superintelligence. 2019-05 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) articulates E05. Closing their review, Trammell and Korinek say that existential risk from AI, central in futurist discussion, is almost absent from economics, because existing growth models offer no obvious mechanism for it. The authors attribute the profession's inattention to growth singularities to a modelling norm tied to the Kaldor facts, not to the models themselves. Footnote dating Sandberg's 2013 overview of singularity models to before mainstream economists' engagement with transformative AI. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI Trammell and Korinek, “Economic Growth under Transformative AI” (2023-10) articulates E05. Cites Jones (2024, 2025) and Shulman and Thornley (2025), and Trammell and Aschenbrenner (2026) on precaution possibly increasing long-term risk. The authors introduce their synthesis of 'the most influential papers on the topic' of transformative AI and growth. Footnote: most models Sandberg summarised did not spell out how AI would fit standard economic models; 'Our review fills this gap.'. 2023-10 · wording attested 2026-04 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Anton Korinek; Donghyun Suh · Scenarios for the Transition to AGI Korinek and Suh, “Scenarios for the Transition to AGI” (2024-03) articulates E05. Final list of next steps after a 'compute-centric' model of the transition to AGI. 2024-03 Inspect claim & evidence ↗ Document page → Articulates · Tier 1 Charles I. Jones · How Much Should We Spend to Reduce A.I.'s Existential Risk? Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E05. Jones then says the Covid-19 experience supplies a key insight, letting existential risk be modelled like pandemic mortality. 2025-03 Inspect claim & evidence ↗ Document page → Texts that bear on it without meeting the criterion (14) Bears on · Tier 1 Robin Hanson · Is a Singularity Just Around the Corner? What it takes to get explosive economic growth Hanson, “Is a Singularity Just Around the Corner? What it takes to get explosive economic growth” (1998-06) bears on E05. Hanson closes by asking singularity advocates to state their case in economic terms. Missing: No statement that economics has neglected or taken up the topic; it sets a standard for advocates. 1998-06 · wording attested 1999-01-28 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Robin Hanson · Economic Growth Given Machine Intelligence Hanson, “Economic Growth Given Machine Intelligence” (2000-08-19) bears on E05. Hanson justifies his paper by a gap in economic research. Hanson's closing call for more economic research. Missing: (E05-02) The neglected topic is machine intelligence replacing labour, not superintelligence, RSI or existential risk. (E05-03) Topic is machine intelligence as labour substitute, not superintelligence, RSI or existential risk. 2000-08-19 · wording attested 2000-08-19 · claimed 1998 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Setting the Stage Hanson, “Setting the Stage” (2008-11-18), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E05. Hanson states that his position in the foom debate rests on growth economics. Missing: Describes Hanson's own method, not whether economics engaged with the topic. 2008-11-18 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · What Core Argument? Hanson, “What Core Argument?” (2008-12-10), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E05. Hanson contrasts Yudkowsky's foom (about sixty doublings in about a week) with specialist estimates. Missing: An unsourced assertion about specialists' likely estimates, not a record of economists' engagement. 2008-12-10 Inspect claim & evidence ↗ Document page → Bears on · Tier 2 Robin Hanson · Foom Debate, Again Hanson, “Foom Debate, Again” (2013-02-18), in “The Hanson-Yudkowsky AI-Foom Debate”, bears on E05. Hanson argues that growth economics, not newly invented categories, should frame the intelligence-explosion question. Missing: Says economics should frame the topic; does not say whether it has. 2013-02-18 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Ari Hyytinen · Ovatko hyvät johtamismenetelmät talouskasvun lähde? (pääkirjoitus) Hyytinen, “Ovatko hyvät johtamismenetelmät talouskasvun lähde? (pääkirjoitus)” (2017) bears on E05. Editorial in the Finnish economics association's journal; it calls the idea "apparently improbable but interesting" before turning to management practices. Missing: Concerns a growth singularity; no statement about economics' engagement with superintelligence or RSI. 2017 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Tyler Cowen · Neglected Open Questions in the Economics of Artificial Intelligence Cowen, “Neglected Open Questions in the Economics of Artificial Intelligence” (2019-05) bears on E05. Cowen's chapter on neglected questions sets aside labour substitution and turns to consumer surplus, trade and redistribution. Missing: No mention of superintelligence, AGI, RSI or AI existential risk in a chapter on neglected questions. 2019-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Philip Trammell; Anton Korinek · Economic Growth under Transformative AI (October 2023 version) Trammell and Korinek, “Economic Growth under Transformative AI (October 2023 version)” (2023-10) bears on E05. The review opens by dating economists' awareness of AI's transformative potential to Simon's prediction. Missing: A claim about economists' long-standing understanding, not a statement that the topic is new or neglected; it qualifies an 'only recently' reading. 2023-10 · wording attested 2023-11-29 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Charles I. Jones · The A.I. Dilemma: Growth versus Existential Risk Jones, “The A.I. Dilemma: Growth versus Existential Risk” (2023-11) bears on E05. Distinguishes the paper from Jones (2016) on growth versus lives lost; earlier the section lists Joy, Bostrom, Rees, Posner, Yudkowsky et al. and Ord as having highlighted AI existential risk. Missing: States the paper's own focus relative to the author's earlier work; it does not say that economics neglected the topic. 2023-11 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Anton Korinek; Donghyun Suh · Scenarios for the Transition to AGI Korinek and Suh, “Scenarios for the Transition to AGI” (2024-03) bears on E05. Followed by Good (1965) on an intelligence explosion, Bostrom (2014) on superintelligence and Yudkowsky (2013) on 'several of the economic implications of the transition to AGI'. Literature review entry on Jones's existential-risk paper. Missing: (E05-23) Connects an economics paper to the superintelligence literature by citation; citation is not engagement with its claims, and the cited authors are non-economists. (E05-24) Acknowledges existential risk as part of AGI discussion; the paper itself does not model it. 2024-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Daron Acemoglu · The Simple Macroeconomics of AI Acemoglu, “The Simple Macroeconomics of AI” (2024-05) bears on E05. First of four possible medium-run macro effects of AI; the paper analyses the more modest task-level effects. Missing: Explicitly sets the singularity aside from the analysis; it neither states nor rejects AI self-improvement or superintelligence, and the 'singularity' meant is a productivity revolution. 2024-05 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Charles I. Jones · How Much Should We Spend to Reduce A.I.'s Existential Risk? Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) bears on E05. Related-literature paragraph that also cites Aschenbrenner (2020), Aschenbrenner and Trammell (2024) and Martin and Pindyck. Missing: Says much of the work on AI existential risk lies outside economics; it is a remark on a review, not a statement that economics neglected the topic. 2025-03 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Erik Brynjolfsson; Anton Korinek; Ajay Agrawal · A Research Agenda for the Economics of Transformative AI Brynjolfsson, Korinek and Agrawal, “A Research Agenda for the Economics of Transformative AI” (2025-09) bears on E05. Opening sentence of the research-agenda abstract ('As we approach Transformative Artificial Intelligence (TAI), …'). The agenda's list; 'safety risks' is one of nine challenges. Missing: (E05-29) Calls for an economics of transformative AI in general; the abstract does not name superintelligence, AGI, the singularity or existential risk. (E05-30) "Safety risks" in an abstract list; existential risk is not stated in the retrieved text. 2025-09 Inspect claim & evidence ↗ Document page → Bears on · Tier 1 Isaiah Andrews; Maryam Farboodi · Do Markets Believe in Transformative AI? Andrews and Farboodi, “Do Markets Believe in Transformative AI?” (2025-09) bears on E05. Event study of US Treasury and TIPS yields around 30 frontier-lab model releases (Nov 2022–Dec 2025). Missing: Evidence about investors' beliefs on transformative AI, inferred from yields; not about superintelligence, self-improvement or existential risk, and not a statement about economics as a field. 2025-09 · wording attested 2026-07 Inspect claim & evidence ↗ Document page →