International network refocuses on advanced AI measurement and evaluation science
Governments participating in an international AI cooperation network have agreed to strengthen their joint work on measuring and evaluating advanced artificial intelligence systems. The group has also confirmed a change in mandate and name, reflecting the growing importance of robust, science-based approaches to AI assessment.
An international group of governments has agreed to deepen cooperation on the science of artificial intelligence measurement and evaluation, following a meeting held in San Diego in early December. The discussions confirmed a shift in the network’s focus and structure, aimed at ensuring that methods for assessing advanced AI systems keep pace with rapid technological development.
Formerly known as the International Network of AI Safety Institutes, the group has been renamed the International Network for Advanced AI Measurement, Evaluation and Science. The change reflects a longer-term emphasis on developing rigorous, comparable, and scientifically grounded approaches to evaluating AI capabilities and risks, rather than a narrower focus on institutional safety frameworks.
As part of the update, the United Kingdom has assumed the role of Network Coordinator. In this capacity, it will support coordination among members and help guide work on shared methodologies for AI measurement and evaluation across jurisdictions.
The network was established in November 2024 and brings together representatives from Australia, Canada, the European Union, France, Japan, Kenya, the Republic of Korea, Singapore, the United Kingdom, and the United States. Since its creation, participants have exchanged technical knowledge and policy experience to align on best practices for evaluating advanced AI systems.
The latest meeting took place alongside the NeurIPS conference, a major international forum for AI research. Holding discussions in parallel with the conference allowed for closer engagement between government officials, academic researchers, and industry experts, and provided an opportunity to connect policy coordination with ongoing scientific work.
Network members have framed improved AI measurement and evaluation as a prerequisite for public trust and informed policymaking. By developing shared scientific foundations for assessing AI capabilities and limitations, the group aims to support the responsible deployment of advanced systems while enabling international cooperation in an area where technologies and impacts increasingly cross national borders.
