The International Network of AI Safety Institutes was launched on 20 November 2024 in San Francisco, at an inaugural convening co-hosted by the US Department of Commerce and the US Department of State. The network drew together the national bodies that governments had begun standing up to test and evaluate advanced AI systems, giving them a shared venue to coordinate technical work across borders.
The ten initial members were Australia, Canada, the European Union, France, Japan, Kenya, the Republic of Korea, Singapore, the United Kingdom, and the United States, with the US AI Safety Institute serving as the inaugural chair. The network’s mission statement set out four priorities: pursuing AI safety research, developing best practices for model testing and evaluation, working toward common approaches for interpreting test results, and advancing global inclusion and information sharing.
The first meeting was structured as a working session on three urgent topics that benefit from international coordination: managing the risks from synthetic content, testing foundation models, and conducting risk assessments for advanced AI systems. The convening was timed to build momentum ahead of the AI Action Summit held in Paris in February 2025, knitting together the summit diplomacy of the Bletchley and Seoul meetings with the more technical, institute-level work of actually evaluating models.