Claim and evidence Close ×
Shah, Varma, Kumar, Phuong, Krakovna, Uesato and Kenton's paper of October 2022 states that a learned system may competently pursue an undesired goal that gives good performance in training, even when the specification is correct, and calls this goal misgeneralization. Copy claim link
Copy claim reference
Machine-checked, not reviewed by an editor Stated in source Supported
Supporting evidence: primary · 1 source; Context: primary · 1 source
Date 2022-10-04 · publication Records dated 2022 → Role Articulation, Earliest located use of the name Register (kind of text) Nonfiction Name used “goal misgeneralization” Scope and limits Records the paper's definition, its examples in current deep-learning systems and its hypotheticals about more capable systems; the authors describe the hypotheticals as speculative and state that how likely the failure is in practice needs further work.
The evidence 03 primary source Supports the claim
Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals ↗ arXiv preprint 2210.01790 (DeepMind) · 2022-10-04
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 “an AI system may pursue an undesired goal even when the specification is correct, in the case of goal misgeneralization” Location in source Abstract, PDF p. 1
How it was read Retrieved directly from the publisher
Machine-checked · 2026-10-31
Source details and check notes Authors Rohin Shah; Vikrant Varma; Ramana Kumar; Mary Phuong; Victoria Krakovna; Jonathan Uesato; Zac Kenton Version Retrieved for the candidate pack on 2026-10-29 (SHA-256 af2ecc2079a4dfcc0a59f7f6f9866ba660fffc073f2078a097f8b1860cd9f134). arXiv v1 of 4 October 2022. Canonical address Open the canonical source ↗ Accessed 2026-10-29 Language English (en) Attribution Rohin Shah, Vikrant Varma, Ramana Kumar, Mary Phuong, Victoria Krakovna, Jonathan Uesato, Zac Kenton, “Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals”, 2022-10-04. 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, arguments about catastrophic risk from advanced AI (research/risk-debate). The quotation (20 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review. primary source Supports the claim
Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals ↗ arXiv preprint 2210.01790 (DeepMind) · 2022-10-04
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 “the learned program competently pursues an undesired goal that leads to good performance in training situations but bad performance in novel test situations” Location in source Abstract, PDF p. 1
How it was read Retrieved directly from the publisher
Machine-checked · 2026-10-31
Source details and check notes Authors Rohin Shah; Vikrant Varma; Ramana Kumar; Mary Phuong; Victoria Krakovna; Jonathan Uesato; Zac Kenton Version Retrieved for the candidate pack on 2026-10-29 (SHA-256 af2ecc2079a4dfcc0a59f7f6f9866ba660fffc073f2078a097f8b1860cd9f134). arXiv v1 of 4 October 2022. Canonical address Open the canonical source ↗ Accessed 2026-10-29 Language English (en) Attribution Rohin Shah, Vikrant Varma, Ramana Kumar, Mary Phuong, Victoria Krakovna, Jonathan Uesato, Zac Kenton, “Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals”, 2022-10-04. 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, arguments about catastrophic risk from advanced AI (research/risk-debate). The quotation (23 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review. primary source Context only
Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals ↗ arXiv preprint 2210.01790 (DeepMind) · 2022-10-04
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 “we would also like to see additional work investigating how likely goal misgeneralization is in practice” Location in source Section 6, PDF p. 12
How it was read Retrieved directly from the publisher
Machine-checked · 2026-10-31
Source details and check notes Authors Rohin Shah; Vikrant Varma; Ramana Kumar; Mary Phuong; Victoria Krakovna; Jonathan Uesato; Zac Kenton Version Retrieved for the candidate pack on 2026-10-29 (SHA-256 af2ecc2079a4dfcc0a59f7f6f9866ba660fffc073f2078a097f8b1860cd9f134). arXiv v1 of 4 October 2022. Canonical address Open the canonical source ↗ Accessed 2026-10-29 Language English (en) Attribution Rohin Shah, Vikrant Varma, Ramana Kumar, Mary Phuong, Victoria Krakovna, Jonathan Uesato, Zac Kenton, “Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals”, 2022-10-04. 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, arguments about catastrophic risk from advanced AI (research/risk-debate). The quotation (16 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms, so the source’s typography can differ. No human editorial review. 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 meets the proposition’s criterion. The record documents what the text states, in its register (fiction and satire are labelled); it is not evidence that the proposition is true and implies no influence on or from another text.
Date. Date of the arXiv v1 submission.
Record ID rd-r03-shah-2022-goal-misgeneralization