Every AI Policy Ever · open in the interactive register
Published (non-binding, current)Non-binding · Voluntary code / commitments
Frontier AI Safety Commitments, AI Seoul Summit 2024
Record INT-06
- Issued by
- AI organisations that agreed to the commitments (20 by February 2025), announced by the UK and Republic of Korea governments at the AI Seoul Summit; published by the UK Department for Science, Innovation and Technology
- Jurisdiction
- AI Seoul Summit (2024)
- Level · region
- International · Global
- Published / adopted
- 2024-05-21
- Last updated
- 2025-02-07
- Language
- English
Summary
Leading AI developers commit to assess frontier-model risks across the lifecycle and to set and monitor thresholds for intolerable severe risks. They agree not to develop or deploy a model if those risks cannot be kept below the thresholds, and to maintain internal accountability. They also commit to publish a safety framework and report publicly on implementation, with deeper information shared with trusted actors such as home governments.
Safety topics
general-purpose / frontier AIcatastrophic riskevaluation & testingtransparency & labellingaccountability & liability
Scope
Borderline scope Company commitments brokered and published by governments; existing scope_note kept.
Status notes
The UK and Republic of Korea governments announced the first 16 signatories: Amazon, Anthropic, Cohere, Google, G42, IBM, Inflection AI, Meta, Microsoft, Mistral AI, Naver, OpenAI, Samsung Electronics, Technology Innovation Institute, xAI and Zhipu.ai. Magic, Minimax, 01.ai and NVIDIA were added on 7 February 2025, making 20. Signatories undertook to publish a safety framework focused on severe risks by the next AI summit, in France (the AI Action Summit, INT-01), and to provide public updates if their approaches to the three outcomes change.
Local copy
Archived. Last attempt 2026-10-17 16:29:34 UTC
Official source (third-party copy, not published) archive/INT-06/2026-10-11/official.html
Retrieved 2026-10-11 13:56:55 UTC
SHA-256 b1f8a2846196606a728e5e831a74ed36e275460a4bbcb68e43ca64f3e663d236
Secondary source (third-party copy, not published) archive/INT-06/2026-10-11/secondary.html
Retrieved 2026-10-11 13:57:03 UTC
SHA-256 4c9c5bc59457b52e34c80e2349ebc23b875e8e621c45719539e2be171cff81d4
Verification
AThe instrument's own official text, or an official register, was stored and confirms the title, date and status. checked 2026-10-17 · official-url-fetched
Stored and read: the GOV.UK publication page (official), which gives the first publication on 21 May 2024 and the update of 7 February 2025 that added organisations, and the commitments' HTML text (alt-official), which gives the 20 organisations, the eight commitments under three outcomes and the definitions; both stored 11 Oct 2026 and unchanged on 17 Oct 2026, with no update after 7 February 2025.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-17: 7 field(s) supported, fixes applied to scope_note.
History
- 2026-10-18related: [{"slug": "seoul-declaration", "id": "INT-04"}, {"slug": "seoul-ministerial-stat → [{"slug": "seoul-declaration", "id": "INT-04"}, {"slug": "seoul-ministerial-stat — two-way link added: GB-02 link(s) to INT-062026-10-18-fix-related-links · Claude Code (claude-opus-5-5)
- 2026-10-17scope_note: Commitments made by AI companies, not by governments; included because governmen → Company commitments brokered and published by governments; existing scope_note k — pipeline run 2026-10-17-recheck-international-2 (recheck), candidate INT-062026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-17notes: The original 16 signatories were Amazon, Anthropic, Cohere, Google, G42, IBM, In → The UK and Republic of Korea governments announced the first 16 signatories: Ama — 'before the Paris AI Action Summit (Feb 2025)' replaced by the text's 'by the upcoming AI Summit in France'; the public-updates clause added2026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-17issuing_body: UK Department for Science, Innovation and Technology (published for the AI Seoul → AI organisations that agreed to the commitments (20 by February 2025), announced — the text says the UK and Republic of Korea governments announced the commitments, which the AI organisations made; DSIT published them. Version 1's 'published for the AI Seoul Summit co-hosts' is replaced by what the page says2026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-14scope_note: – → Commitments made by AI companies, not by governments; included because governmen — borderline under the include-and-flag rules of issue #22026-10-14-issue-2-scope-notes · Claude Code (claude-opus-5-5)
- 2026-10-14notes: The original 16 signatories were Amazon, Anthropic, Cohere, Google, G42, IBM, In → The original 16 signatories were Amazon, Anthropic, Cohere, Google, G42, IBM, In — scope caveat moved to scope_note2026-10-14-issue-2-scope-notes · Claude Code (claude-opus-5-5)
Citation
AI organisations that agreed to the commitments (20 by February 2025), announced by the UK and Republic of Korea governments at the AI Seoul Summit; published by the UK Department for Science, Innovation and Technology. Frontier AI Safety Commitments, AI Seoul Summit 2024. 2024-05-21. https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/INT-06/2026-10-11/official.html retrieved 2026-10-11 13:56:55 UTC, SHA-256 b1f8a2846196606a728e5e831a74ed36e275460a4bbcb68e43ca64f3e663d236. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record INT-06.
Improve this record
Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.