AI Safety Summit held at Bletchley Park

Britain has hosted the first global summit on the safety of artificial intelligence at Bletchley Park, the Buckinghamshire estate where codebreakers worked during the Second World War.

The two-day summit, which opened on 1 November, brought together representatives from 28 countries, including the United States and China, along with technology companies and experts.

The Bletchley Declaration

On the first day, the countries attending signed the Bletchley Declaration, agreeing that the most advanced AI systems, known as frontier AI, could pose serious and even catastrophic risks, and that countries must work together to understand and manage them.

The declaration was signed by the United States, China and the European Union, as well as Britain and other countries. The Prime Minister, Rishi Sunak, said it was a landmark achievement to get such a wide range of countries to agree.

Testing new models

Several leading AI companies agreed to allow governments to test their most advanced models before they are released. Britain has set up an AI Safety Institute to carry out such testing, and the United States announced its own institute.

Further summits are planned, with South Korea and France agreeing to host the next meetings.

Who was there?

Those attending included the American vice-president, Kamala Harris, the president of the European Commission, Ursula von der Leyen, and the United Nations Secretary-General, Antonio Guterres. China sent a vice-minister.

Technology leaders included Elon Musk, who took part in a conversation with Sunak broadcast after the summit, and the heads of leading AI companies, including OpenAI, Google DeepMind and Anthropic.

Why now?

The summit comes after a year of rapid progress in AI, following the release of ChatGPT and other tools. Some experts have warned that powerful AI could be used to create cyber attacks or biological weapons, or could one day escape human control.

Others argue that the focus on long-term risks distracts from more immediate problems, such as bias, misinformation and the effect on jobs.

A British role

Sunak wants Britain to be a world leader in AI safety. The country is home to a large AI industry, including Google DeepMind in London.

Bletchley Park was chosen for its history. It was there that Alan Turing and other codebreakers helped break the German Enigma code, work that played an important part in the development of computers.