AI Safety Summit 2023

• 1. Introduction

  ›

  The AI Safety Summit 2023 opened in November 2023 at Bletchley Park in Milton Keynes, England. It was the first ever global conference dedicated to the safety of artificial intelligence. Rishi Sunak, Prime Minister of the United Kingdom, had made AI a stated priority of his government. He had announced that Britain would host the conference in autumn 2023. Twenty-eight countries participated and produced the Bletchley Declaration. Who was invited to sit at the table? What did those nations commit to? And what measures did governments put in place when they returned home?

  124711

• 2. Where Computers Were Born

  ›

  The British government established Bletchley Park as a codebreaking facility during the Second World War. Its buildings and grounds had been developed around a Victorian manor. Within those grounds, some of the first modern computers were constructed. That codebreaking work gave the facility a lasting role in the history of computing. The facility that had broken enemy codes in wartime would now be the setting for the first international attempt to write rules for artificial intelligence.

  817

• 3. Shared Commitments

  ›

  The Bletchley Declaration called for artificial intelligence to be developed, deployed, and used "in a manner that is safe, human-centric, trustworthy and responsible." The declaration set out two specific goals: identifying AI safety risks of shared concern, and building risk-based policies in response to those risks. Its stated aim was to ensure "that the benefits of the technology can be harnessed responsibly for good and for all." International co-operation was a central theme. The declaration was not just a statement of values. It was a call for nations to act together in managing a technology whose risks no single country could address alone. Those commitments raised the question of how exactly to identify and govern the most capable AI systems.

  510119

• 4. What Governments Feared

  ›

  The summit introduced the term "Frontier AI" to describe the newest and most capable artificial intelligence systems. These were placed at the center of the declaration's concerns. The risks identified at the summit ranged from the immediate to the catastrophic. AI's potential use for terrorism and criminal activity represented one class of concern. Its potential role in warfare was another. Delegates also raised the possibility of AI posing an existential risk to humanity as a whole. Joe Biden, President of the United States, signed an executive order requiring AI developers to share their safety results with the US government. The US government also announced the creation of an American AI Safety Institute, part of the National Institute of Standards and Technology. The summit's attendees included technology executives from the world's most powerful AI companies alongside heads of state and senior government officials.

  41415612

• 5. In the Room

  ›

  Kamala Harris, Vice President of the United States, represented the Biden administration at the summit. Giorgia Meloni, Prime Minister of Italy, attended alongside Ursula von der Leyen, President of the European Commission. Wu Zhaohui, China's vice-minister of science and technology, was also present. King Charles III attended virtually. Sam Altman, CEO of OpenAI, and Mustafa Suleyman, co-founder of DeepMind, were among the technology executives in attendance. Nick Clegg, a former British politician, was at the time president of global affairs at Meta Platforms. Gina Raimondo, United States Secretary of Commerce, attended alongside Michelle Donelan, the UK Secretary of State for Science, Innovation and Technology. Vera Jourova, vice-president of the European Commission for Values and Transparency, was also in the room. Elon Musk, CEO of Tesla and owner of X, SpaceX, Neuralink, and xAI, also attended. On the 2nd of November, Musk held a live conversation with Rishi Sunak on X about artificial intelligence safety. The summit began a series that would see five countries, across Europe and Asia, each take a turn as host.

  1617191815

• 6. A Continuing Series

  ›

  The AI Seoul Summit, held in Seoul, South Korea, followed in 2024. France hosted the AI Action Summit in Paris in 2025. The AI Impact Summit was scheduled for New Delhi, India, in 2026. Geneva, Switzerland, was set to host the AI Summit in 2027. The gathering in Geneva would come four years after the delegates first met at Bletchley Park.

  123