CAICT Launches 2026 AI Safety Evaluations Initiative

The China Academy of Information and Communications Technology (CAICT) has officially launched its 2026 AI safety evaluations. This initiative invites Chinese companies to register their AI models for assessment based on multiple safety and security standards. The evaluations are set occur in June and July 2026, and will focus on key areas such as internal AI safety capabilities, platform trustworthiness, application security, and infrastructure autonomy rates.

This program marks a significant shift in China's commitment to improving AI safety standards. By conducting these evaluations, CAICT aims to address potential risks identified in previous tests, bolstering the overall reliability of domestic AI technologies. This initiative enhances national capabilities in AI safety, significantly reducing reliance on foreign technology and fortifying China's position in the global AI landscape.