The Bletchley Declaration on AI Safety

The Bletchley Declaration on AI Safety was adopted at the inaugural AI Safety Summit, convened at Bletchley Park under UK leadership and joined by 28 countries and the European Union. It represents the first coordinated international political commitment focused specifically on the safety of “frontier AI” systems—those capable of advanced or unpredictable behaviours that may present serious or even catastrophic risks. The Declaration affirms that AI offers vast opportunities for global progress but stresses shared responsibility for addressing safety challenges through international collaboration. It commits participating governments to advance scientific understanding of AI risks, promote transparency and accountability among developers, and support the creation of testing, evaluation, and governance mechanisms. Although non-binding, it establishes a normative foundation for subsequent safety initiatives and multilateral cooperation, linking with later forums such as the Seoul 2024 and Paris 2025 AI Safety Summits.