Spooked by Mythos, Trump Suddenly Realized AI Safety Testing Might Be Good

The Trump administration signed voluntary deals with major AI firms to allow government safety checks on frontier models before and after release.

The shift comes after Anthropic held back Claude Mythos over cybersecurity concerns, underscoring growing national security worries around advanced AI.

Officials said an executive order requiring pre-release testing could follow, marking a possible return to stronger AI oversight.

The move could shape how frontier models are evaluated, regulated, and trusted publicly.