In November 2023, world leaders and technologists specializing in artificial intelligence gathered at Bletchley Park, the British estate that housed Alan Turing’s team of codebreakers during World War II, for the world’s first AI Safety Summit. Convened just 11 months after OpenAI’s launch of ChatGPT thrust AI into the cultural and political spotlight, the meeting was framed as a necessary attempt to confront the catastrophic and even existential risks that increasingly powerful AI systems could potentially pose to society.
What emerged from Bletchley, however, was not an international governance framework but a diplomatic declaration. The summit did produce tangible outcomes, including the establishment of the U.K. AI Safety Institute—now renamed as the AI Security Institute—and the creation of an expert panel to generate scientific reporting on frontier AI. But the harder questions of enforcement and international accountability were left largely unresolved.
A follow-up summit in Seoul in May 2024 began to dilute the original safety focus by folding in innovation and inclusivity as additional core issues. By the time of the AI Action Summit in Paris in February 2025, the gatherings had largely ceased to function as forums for governing AI risk and instead became more of a venue for promoting accelerated deployment and adoption, while loosely acknowledging potential harms.