Trump's AI policy does a major U-turn: plans to establish a pre-release review mechanism, with Anthropic's unreleased model becoming a trigger.

According to Beating Monitoring, the Trump administration, which previously advocated a non-interventionist stance on AI, is changing its policy and is discussing establishing an AI task force through an executive order, aiming to introduce a government review mechanism for new AI model releases.

The policy shift is directly triggered by Anthropic’s launch last month of a new model called Mythos. This model has a very strong ability to identify software security vulnerabilities and is believed to potentially trigger a “cybersecurity purge.” After assessing the risks, Anthropic refused to release it to the public. To prevent catastrophic AI cyberattacks from causing political backlash, some U.S. officials are pushing to establish a review system that requires the government to have “priority access” to new models, but does not prevent their final release.

This policy reversal is currently led by White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent. Behind the scenes, there is a complex supply and demand relationship between the U.S. government and leading AI companies: in March this year, the Pentagon cut off Anthropic’s services over a $200 million contract dispute, and Anthropic subsequently sued the U.S. government. However, the NSA is still using Mythos to evaluate software vulnerabilities in U.S. government systems, and the U.S. military continues to rely on the Maven system, which incorporates Anthropic technology, in its war against Iran. Two senior White House officials met with Anthropic CEO Dario Amodei last month in an attempt to restore full government use of the technology.