The Trump administration is weighing mandatory pre-release vetting of AI models, according to reporting by Tom's Hardware. Anthropic's recent Mythos instance is cited as a catalyst for the reconsideration of the administration's previous hands-off AI policy stance. The exact scope and enforcement mechanism have not been finalized.
If enacted, the policy would require developers to submit models to a government authority for safety review before public or commercial release. The shift reflects broader concern among policymakers about model misuse and safety, and places the U.S. approach closer to EU regulatory frameworks, though details remain under discussion.