Some 28 countries, including major AI nations such as the UK, US, EU, and China, have reached a groundbreaking agreement at Bletchley Park to establish a shared understanding of the risks and opportunities posed by frontier AI. The Bletchley Declaration on AI safety acknowledges the potential for serious harm from intentional misuse or unintended control issues of frontier AI, but also recognises the benefits of this technology for society, such as improved healthcare and education. The agreement calls for sustained international collaboration on frontier AI safety and research, building on existing efforts and supporting a network of scientific research.
At the AI Safety Summit, the UK Government made a number of announcements, including that the UK will establish the world’s first AI Safety Institute and host the next virtual summit, and a £100 million fund to capitalise on AI’s potential in life sciences and healthcare. The UK Prime Minister’s closing speech announced that world leaders have set out a plan for safety testing frontier AI, including the ‘State of Science’ Report to be led and delivered by Yoshua Bengio, widely regarded as the ‘Godfather of AI’.
Relevant background information from the Department for Science, Innovation and Technology can be found on the links below:
- Frontier AI Taskforce: second progress report
- Frontier AI: capabilities and risks – discussion paper
- AI Standards Hub: pilot evaluation report
- Frontier AI Taskforce: first progress report