Catenaa, Wednesday, June 10, 2026- The White House has directed a federal artificial intelligence testing agency to stop releasing public reports on advanced AI model evaluations, citing national security concerns as the administration seeks to balance rapid AI innovation with tighter oversight of frontier technologies.
The move affects the Center for AI Standards and Innovation (CAISI), the federal body responsible for assessing advanced AI systems for risks related to cybersecurity, biosecurity and other national security threats before broader deployment.
The directive does not end government testing of advanced AI models.
Instead, it shifts the results of those evaluations from public disclosure to internal government channels.
CAISI, formerly known as the U.S. AI Safety Institute, continues to work with major AI developers under voluntary agreements designed to evaluate the capabilities and risks of cutting-edge models.
Participating companies include OpenAI, Anthropic, Google DeepMind, Microsoft and xAI.
The policy follows a June 2 executive order that emphasizes maintaining U.S. leadership in artificial intelligence while strengthening protections around advanced technologies.
Administration officials argue that public disclosure of certain testing findings could expose sensitive information about frontier AI systems and create security risks.
Supporters of the decision say limiting access to assessment reports may help prevent misuse of technical information by hostile actors.
Critics, however, warn that reducing transparency could make it harder for investors, researchers and the public to independently evaluate AI safety claims.
Because participation in CAISI’s testing program remains voluntary, some observers argue that public reporting has served as an important accountability mechanism for AI developers.
The change may also reduce the reputational benefits companies receive from demonstrating cooperation with government safety evaluations.
The decision comes as artificial intelligence becomes increasingly intertwined with sectors including finance, cybersecurity and cryptocurrency.
Many blockchain and crypto projects are integrating AI-powered tools and autonomous agents, making public trust and transparency important factors in adoption.
Analysts said investors will closely monitor whether major AI firms continue participating in voluntary testing programs and whether future regulations introduce more formal oversight requirements.
CAISI was established to evaluate advanced artificial intelligence systems for national security and public safety risks. The agency conducts testing in areas such as cyber capabilities, biological threat detection and model reliability. Unlike traditional regulatory bodies, it operates through voluntary cooperation with AI developers rather than mandatory licensing requirements. The latest policy shift reflects a broader debate within Washington over whether AI governance should prioritize transparency or security as increasingly powerful AI systems enter commercial use.
