Textbook of AI

The AI Safety Summit held at Bletchley Park on 1–2 November 2023 was the first international summit specifically devoted to the safety of advanced AI. Convened by the United Kingdom under Prime Minister Rishi Sunak, it brought together twenty-eight participating governments, the European Union and leading AI companies, producing the Bletchley Declaration, a multilateral commitment to international cooperation on identifying and managing the risks of frontier AI.

The location

Bletchley Park was Britain's wartime codebreaking centre and the site where Alan Turing, Gordon Welchman, Tommy Flowers and a workforce of nearly ten thousand broke the German Enigma and Lorenz ciphers, work credited with shortening the Second World War by an estimated two years. The choice of venue signalled that frontier-AI safety was to be treated as a strategic, international and historically resonant issue, not merely a technical one.

The Declaration

The Bletchley Declaration was signed by the United Kingdom, the United States, the European Union, China, India, Japan, Brazil, Australia, Canada, France, Germany, Italy, the Netherlands, Spain and others. It affirmed that "AI should be designed, developed, deployed, and used in a manner that is safe, in such a way as to be human-centric, trustworthy and responsible," and committed signatories to:

• Identifying AI safety risks of shared concern.
• Building a shared scientific and evidence-based understanding of those risks.
• Developing risk-based policies across countries to ensure safety.

The participation of both the United States and China, at a moment of substantial geopolitical tension over semiconductor export controls and Taiwan, was widely noted as significant. It was the first time the two nations had jointly endorsed a multilateral statement on AI risk.

Companion announcements

Alongside the declaration, frontier-AI laboratories, OpenAI, Anthropic, Google DeepMind, Microsoft, Meta, Inflection, Amazon and Mistral , agreed to allow government safety institutes pre-deployment access to their next-generation models. The UK announced the establishment of the UK AI Safety Institute (AISI), the first state body dedicated to evaluating frontier models; the United States followed with the US AI Safety Institute within NIST.

Follow-up

The summit established a series of regular meetings:

• AI Seoul Summit, May 2024, co-hosted by the UK and Republic of Korea, produced the Seoul Declaration and saw companies commit to publish responsible-scaling policies.
• AI Action Summit, Paris, February 2025, hosted by France, broadened the agenda to inclusion, environment and labour.

Significance

The Bletchley summit is generally taken as the inflection point at which advanced-AI governance moved from a niche academic and industry concern to mainstream international policy. It established AI safety institutes as the standard institutional response, normalised pre-deployment government evaluation of frontier models, and set the precedent for treaty-style multilateral cooperation on AI risk. It is often compared, by its organisers, to the early international conferences on nuclear-weapons safeguards.