Pentagon halts Claude AI use, except for nuclear defense
The Pentagon is giving military teams 180 days to stop using Anthropic's Claude AI, but there are exceptions for high-stakes national security tasks like nuclear command, missile defense, and cyber operations.
Only the Pentagon's chief information officer can approve these rare uses, with strict risk checks and notifications for contractors.
President orders ban after Anthropic lawsuit
This move follows President Trump's order to ban Anthropic's AI after the company refused to drop safety limits on surveillance and autonomous weapons.
Anthropic pushed back with a lawsuit, saying the ban goes too far.
Legal experts expect a wave of waiver requests as teams scramble for alternatives, especially since OpenAI quickly struck its own deal with the Pentagon, which the company said included safeguards though the published contract primarily referenced existing laws and policies.
One analyst noted the decision underscored how central advanced AI tools have become to national-security operations.