Howard Lutnick, the U.S. Commerce Secretary, announced a sudden policy reversal on Tuesday as the Trump administration moved to implement safety checks on advanced artificial intelligence. Microsoft, Google’s DeepMind, and Elon Musk’s xAI have agreed to share early versions of their models with the newly rebranded Center for AI Standards and Innovation (CAISI). The move comes after Anthropic announced its “Mythos” model was too dangerous for public release due to its ability to identify and exploit systemic cybersecurity vulnerabilities.

“Mythos is the first, but it’s incumbent on us to build a system so U.S. AI can be the leader in AI and be safe at the same time,” according to a CAISI press release. The administration, which previously campaigned against AI regulation as “burdensome,” is now using CAISI to conduct pre-deployment evaluations. These agreements renegotiate previous Biden-era safety frameworks to align with President Trump’s directives. Microsoft acknowledged that testing for national security risks must be a collaborative endeavor with government institutions like CAISI.

This paper identifies this shift as the consolidation of the Cognitive Enclosure. By gating advanced models like Mythos behind a state-directed security screen, the administration ensures that the most powerful autonomous tools are managed by a corporate-state alliance rather than the public. The transformation of the U.S. Artificial Intelligence Safety Institute into CAISI—stripping the word "safety" from its title—signals that the focus has moved from protecting the public to securing the technological frontier for elite capital and military application.