Every AI Policy Ever · open in the interactive register
Published (non-binding, current)Non-binding · Declaration / statement
The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023
Record INT-03
- Issued by
- Governments attending the AI Safety Summit at Bletchley Park (28 countries and the EU), hosted by the UK Government
- Jurisdiction
- AI Safety Summit (Bletchley, 2023)
- Level · region
- International · Global
- Published / adopted
- 2023-11-01
- Last updated
- 2025-02-13
- Language
- English
Summary
Declaration by 28 countries and the EU on the risks of 'frontier' AI: highly capable general-purpose models, and narrow AI with harmful capabilities, that match or exceed today's most advanced models. It warns of substantial risks from intentional misuse or unintended issues of control, names cybersecurity, biotechnology and disinformation, and sees potential for serious, even catastrophic, harm. Its agenda is a shared scientific understanding of these risks and risk-based national policies, with safety testing, evaluations and transparency from developers of frontier AI, and an international network of scientific research on frontier AI safety.
Safety topics
general-purpose / frontier AIcatastrophic riskevaluation & testingtransparency & labellinginternational cooperationresearch priorities
Status notes
Agreed at the AI Safety Summit at Bletchley Park, 1-2 November 2023; the countries represented include the US, China and the EU. New Zealand joined the commitment on 23 October 2024, and GOV.UK updated the publication on 13 February 2025 to include it as a signatory. The participants said they would meet again in 2024. The International AI Safety Report (interim report INT-08, full report INT-09) was launched at this summit. GOV.UK also offers a Latin American Spanish version of the publication.
Local copy
Archived. Last attempt 2026-10-17 16:29:29 UTC
Official source (third-party copy, not published) archive/INT-03/2026-10-17/official.html
Retrieved 2026-10-17 16:29:35 UTC
SHA-256 dc20cd93d415b40e9ec33b4ad22532afe3fbc2557efc80ed27f8354f9ade8953
Secondary source (third-party copy, not published) archive/INT-03/2026-10-17/secondary.html
Retrieved 2026-10-17 16:29:44 UTC
SHA-256 5ee343f70f6641d37322dcb2a3d10450f46f567d5849ed0543753c46899c736e
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 declaration's HTML text on GOV.UK (official, stored 11 Oct 2026 and unchanged on 17 Oct 2026), which confirms the title, the text, the 28 countries and the EU represented and New Zealand's accession of 23 October 2024; and, in this run, the GOV.UK publication page (proposed as secondary_url), which gives the first publication on 1 November 2023 and the update of 13 February 2025 that added New Zealand, and DSIT's press release of 29 January 2025, which says the International AI Safety Report was launched at the summit. Version 1's claim that the text was unchanged by the 2025 update, and its description as the 'first' declaration of its kind, are not supported and were removed.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-17: 10 field(s) supported.
History
- 2026-10-17notes: Signed at the first AI Safety Summit, 1-2 Nov 2023. Signatories include the US, → Agreed at the AI Safety Summit at Bletchley Park, 1-2 November 2023; the countri — the 13 February 2025 update added New Zealand as a signatory (version 1 said the text was unchanged); New Zealand's accession; the launch of the International AI Safety Report at the summit from DSIT's press release; the later summits are left to the related records2026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-17summary: First intergovernmental declaration focused on 'frontier AI' (highly capable gen → Declaration by 28 countries and the EU on the risks of 'frontier' AI: highly cap — 'First intergovernmental declaration focused on frontier AI' is a claim the text does not make; the definition of frontier AI, the risks named and the agenda are now as the text gives them2026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-17secondary_url: – → https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-d — the GOV.UK publication page (linked from the stored text), stored in this run: the only stored source for the first-publication date and for what the 13 February 2025 update changed2026-10-17-recheck-international-2 · Claude Code (claude-opus-5-5)
- 2026-10-14language: English (translations published) → English — The field held a remark after the language name; it now gives the language names only (AGENTS.md 5.6).manual-2026-10-14 · Claude Code (claude-opus-5-5)
Citation
Governments attending the AI Safety Summit at Bletchley Park (28 countries and the EU), hosted by the UK Government. The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023. 2023-11-01. https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/INT-03/2026-10-17/official.html retrieved 2026-10-17 16:29:35 UTC, SHA-256 dc20cd93d415b40e9ec33b4ad22532afe3fbc2557efc80ed27f8354f9ade8953. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record INT-03.
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.