The Trump administration is considering mandatory reviews of AI models to address security risks.

Techub News — According to CryptoBriefing, the Trump administration is considering mandatory security reviews for new AI models, requiring them to pass government assessments before public release. This move stems from Anthropic's Mythos model, which has demonstrated the ability to discover hidden software vulnerabilities with national security implications.

The administration, previously favoring a deregulatory stance, has recently held discussions with executives from Anthropic, Google, and OpenAI on AI safety. Analysts note that if U.S. AI models can identify critical system vulnerabilities, rival nations may also possess similar capabilities, making reviews necessary to prevent the weaponization of the technology.