Microsoft Expands AI Evaluation Efforts Through U.S. and U.K. Partnerships
Microsoft announced expanded collaboration with the Center for AI Standards and Innovation (CAISI) in the United States and the AI Security Institute (AISI) in the United Kingdom to advance evaluation methods for artificial intelligence systems. The effort focuses on developing approaches for testing, measuring, and improving AI safety, reliability, and security. The partnerships are intended to support research into evaluating AI models and establishing practices that can be used across industry and government. These collaborations are designed to improve measurement science, evaluation methodologies, practical testing workflows, and real-world mitigation impact.
The work will include collaboration on benchmarking, threat analysis, and assessment methodologies designed to address risks associated with increasingly capable AI systems. The initiative also reflects growing international coordination around AI governance and standards development. The expansion highlights how Microsoft is bringing technical expertise and operational experience to strengthen AI evaluation methods and practical testing foundations, testing Microsoft's frontier models to mitigate security risks towards a safe and trustworthy system.
Posted by Pure AI Editors on 05/11/2026