Choose an institution to see its records and dossier, or choose a topic.
Through year 2027 · all records, including scheduled dates 2026 2025 2024 2023 2022 2021 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2008 2005 1965 Show Organisations and infrastructure All types Organisation Person Document Concept Infrastructure Public body Reset filters
94 people, organisations and other records · 122 relationships · 34 claims about a valueEverything recorded up to the chosen year, whether or not still in effect · Undated records included
Claim and evidence Close ×
Did RSP v2.0’s changes to threshold and mitigation specification reduce the verifiability of Anthropic’s safety commitments relative to v1.0? Copy claim link
Copy claim reference
Machine-checked, not reviewed by an editor Question posed by Selekoa Unresolved
Unresolved. The recorded evidence does not settle this question. The linked evidence is background only: it neither supports nor contradicts the claim.
Context: primary + secondary · 2 sources
Basis of the change Assessed effect. An open question, and not a finding. Subject version v2.0 · Publication date unknown · Effective 2024-10-15 Object version v1.0 · Published 2023-09-19 · Effective 2023-09-19 Date 2024-10-15 · effective Scope and limits Unresolved open question, not a finding. SaferAI characterises the revision as replacing precise thresholds and mitigation specifications with qualitative descriptions; Anthropic describes greater flexibility and retained core safeguards. These passages do not establish opposing assessments of verifiability. SaferAI judges that the change reduces transparency and accountability and asks stakeholders to trust Anthropic rather than providing verifiable commitments and metrics, and it lowered Anthropic’s rating from 2.2 to 1.9; the record holds this criticism but no direct opposing assessment of verifiability, so every passage is context. The textual changes are recorded separately, and the record does not decide the question.
The evidence 03 secondary source Context only
Anthropic’s responsible scaling policy update makes a step backwards ↗ SaferAI · 2024-10-23
Archived copy, captured 2024-11-09: Anthropic’s responsible scaling policy update makes a step backwards ↗
Copy source reference
Cite the original source Cite the original document, not Selekoa. Check these details against it.
Copy APA
Download BibTeX Download RIS Download CSL-JSON
Source excerpt “Anthropic’s responsible scaling policy update makes a step backwards” Location in source Headline
Machine-checked · 2026-10-03
Source details and check notes Authors SaferAI Version Dated SaferAI blog post assessing Anthropic’s October 2024 RSP update. It was first published at www.safer-ai.org/post/anthropics-responsible-scaling-policy-update-makes-a-step-backwards and moved to its current address by 2026-02-03, when the Internet Archive recorded a redirect. The archive link is the capture of 2024-11-09 at the earlier address, 17 days after publication; its text matches the live page word for word apart from quotation-mark style. The headline was in title case at publication and is now in sentence case. Quotations match the live page and the captures of 2024-11-09 and 2026-02-03. Canonical address Open the canonical source ↗ Accessed 2026-10-04 Language English (en) Archive copy Open archived copy, captured 2024-11-09: Anthropic’s responsible scaling policy update makes a step backwards ↗ . Captured 17 days after the recorded publication date. An archive copy shows only what the archive retrieved on 2024-11-09. It does not prove what the page said earlier, and it can differ from the copy checked on 2026-10-04.Attribution SaferAI, Anthropic’s responsible scaling policy update makes a step backwards (2024-10-23). Reuse No open redistribution licence asserted. Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice. Check notes Verbatim excerpt checked by an automated research assistant on 2026-10-03 against the live page. The Internet Archive captures of 2024-11-09 (at the post’s earlier address) and 2026-02-03 show the same headline in title case, its form at publication; the live page now sets it in sentence case. SaferAI’s overall judgement of the update. Secondary commentary: evidence of the assessment, not of what the policy text says. Recorded as context on 2026-10-04: the record is an open question, and no passage in it directly opposes an assessment of verifiability. Human editorial review is pending. secondary source Context only
Anthropic’s responsible scaling policy update makes a step backwards ↗ SaferAI · 2024-10-23
Archived copy, captured 2024-11-09: Anthropic’s responsible scaling policy update makes a step backwards ↗
Copy source reference
Cite the original source Cite the original document, not Selekoa. Check these details against it.
Copy APA
Download BibTeX Download RIS Download CSL-JSON
Source excerpt “This shift significantly reduces transparency and accountability, essentially asking stakeholders to accept a “trust us to handle it appropriately” approach rather than providing verifiable commitments and metrics.” Location in source Second paragraph, final sentence
Machine-checked · 2026-10-04
Source details and check notes Authors SaferAI Version Dated SaferAI blog post assessing Anthropic’s October 2024 RSP update. It was first published at www.safer-ai.org/post/anthropics-responsible-scaling-policy-update-makes-a-step-backwards and moved to its current address by 2026-02-03, when the Internet Archive recorded a redirect. The archive link is the capture of 2024-11-09 at the earlier address, 17 days after publication; its text matches the live page word for word apart from quotation-mark style. The headline was in title case at publication and is now in sentence case. Quotations match the live page and the captures of 2024-11-09 and 2026-02-03. Canonical address Open the canonical source ↗ Accessed 2026-10-04 Language English (en) Archive copy Open archived copy, captured 2024-11-09: Anthropic’s responsible scaling policy update makes a step backwards ↗ . Captured 17 days after the recorded publication date. An archive copy shows only what the archive retrieved on 2024-11-09. It does not prove what the page said earlier, and it can differ from the copy checked on 2026-10-04.Attribution SaferAI, Anthropic’s responsible scaling policy update makes a step backwards (2024-10-23). Reuse No open redistribution licence asserted. Selekoa practice: metadata, attributed paraphrases and short attributed quotations under the editorial policy (at most 50 words each; at most 400 quoted words per document across the dataset, and about 10% of a document under about 2,000 words). Documents are linked externally, not redistributed. Publisher terms that forbid quotation override this practice. Check notes Verbatim excerpt checked by an automated research assistant on 2026-10-04 against the live page and the Internet Archive captures of 2024-11-09 (earlier address; straight quotation marks) and 2026-02-03. The same post reports that SaferAI lowered Anthropic’s rating from 2.2 to 1.9. Secondary commentary: evidence of the assessment, not of what the policy text says. Recorded as context on 2026-10-04: the record is an open question, and no passage in it directly opposes an assessment of verifiability. Human editorial review is pending. primary source Context only
Announcing our updated Responsible Scaling Policy ↗ Anthropic · 2024-10-15
Archived copy, captured 2024-10-20: Announcing our updated Responsible Scaling Policy ↗
Copy source reference
Cite the original source Cite the original document, not Selekoa. Check these details against it.
Copy APA
Download BibTeX Download RIS Download CSL-JSON
Paraphrase Anthropic presents the update as a more flexible, nuanced way of assessing and managing AI risks that keeps its commitment not to train or deploy models without adequate safeguards, and names new capability thresholds, refined assessment processes and new internal-governance and external-input measures as key improvements.
Location in source Opening paragraph
Machine-checked · 2026-10-04
Source details and check notes Authors Anthropic Version Page inspected 2026-09-19; historical statements are scoped to their source, not treated as current appointments. Canonical address Open the canonical source ↗ Accessed 2026-09-19 Language English (en) Archive copy Open archived copy, captured 2024-10-20: Announcing our updated Responsible Scaling Policy ↗ . Captured 5 days after the recorded publication date. An archive copy shows only what the archive retrieved on 2024-10-20. It does not prove what the page said earlier, and it can differ from the copy checked on 2026-09-19.Attribution Anthropic, Announcing our updated Responsible Scaling Policy (2024-10-15) Reuse No open redistribution licence asserted. Bibliographic metadata and brief attributed paraphrases only; no full source or personal contact details redistributed. Check notes Paraphrase (the source’s recorded reuse note permits paraphrase only) checked by an automated research assistant on 2026-10-04 against the live page and the Internet Archive capture of 2024-10-20, whose article text is identical. Reworded on 2026-10-04 to name the three changes the opening paragraph calls key improvements and to avoid reusing the source’s wording at length. The publisher’s characterisation of the same shift, not independent evidence of effect. Recorded as context on 2026-10-04: retaining the core safeguard commitment and calling the changes improvements do not deny reduced verifiability. Human editorial review is pending. Status and meaning Editorial status. Not reviewed by an editor . An exploratory record that no editor has accepted. This status does not say whether a person has read any linked passage. Each evidence record above shows how it was checked.
Relation. One scoped difference between document versions. The claim identifies publisher-stated change, official changelog/redline, or an explicitly inferred comparison. It does not establish implementation, effectiveness or a safety ranking.
Date. v2.0’s effective date; the assessments were published afterwards (SaferAI on 2024-10-23).
Record ID rsp-v2-0-assessed-weakening