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 ×
“Creating Friendly AI” (2001-06-15) bears on Y25. A table cell giving the design assumption that is conservative for Friendly AI. It stands opposite the nonconservative assumption 'Human-observable cognition; AI can be definitely known to be Friendly.' Copy claim link
Copy claim reference
Machine-checked, not reviewed by an editor Stated in source Supported
Supporting evidence: primary · 1 source
Date 2001-06-15 · publication Wording attested Only in a later, undated copy. Known only from MIRI's later, undated PDF reprint; the wording was not found in the 2001 HTML pages compared. Role Context Register (kind of text) Nonfiction Missing element A worst-case design assumption that an AI could hide unfriendly cognition; it does not address whether reading an AI's thoughts makes it safe. Scope and limits Premise addressed: We can see what an AI system is thinking. Creating Friendly AI is dated 15 June 2001; the cached text is a later MIRI-hosted PDF reprint, and table cells were reconstructed from extracted PDF text. The same document proposes a 'flight recorder' so that an AI's past states can be inspected, so it treats inspection as a safeguard while not assuming it is sufficient. A worst-case design assumption from a 2001 table, contrasted with "human-observable cognition"; it predates modern interpretability research.
The evidence 01 primary source Supports the claim
Creating Friendly AI 1.0 (MIRI reprint) ↗ Singularity Institute for Artificial Intelligence · 2001-06-15
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 “Opaque cognition; the AI would probably succeed in hiding unFriendly cognition if it tried.” Location in source Creating Friendly AI 1.0, section 2.2, table 'Additional Assumptions', column 'Conservative for Friendly AI', row on cognition, printed p. 9
Machine-checked · 2026-10-06
Source details and check notes Authors Eliezer Yudkowsky Version Singularity Institute, MIRI reprint, retrieved 2026-10-06 (SHA-256 51f21b8b70c2b7324c20de85edea40b2ef224cadff56ef7e0a3aa2a08dba2752). Release date from the reprint; an open-commentary version was announced on SL4 on 2001-04-18, so some wording may be earlier than this date. Reprint wording is not authenticated for 2001 unless a capture was compared. Canonical address Open the canonical source ↗ Accessed 2026-10-06 Language English (en) Attribution Eliezer Yudkowsky, “Creating Friendly AI 1.0 (MIRI reprint)”, Singularity Institute, MIRI reprint, 2001-06-15. Reuse No open redistribution licence asserted. Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution. Check notes Candidate pack passage T25-a. The quotation (14 words) matched the retrieved expression (SHA-256 51f21b8b70c2b7324c20de85edea40b2ef224cadff56ef7e0a3aa2a08dba2752) when checked on 2026-10-06. 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. A passage of the document bears on the proposition without meeting its criterion; the record names the missing element. It is not an articulation and implies no influence.
Date. Release date from the reprint; an open-commentary version was announced on SL4 on 2001-04-18, so some wording may be earlier than this date. Reprint wording is not authenticated for 2001 unless a capture was compared.
Record ID yk-y25-creating-friendly-ai-2001-b