Pentagon ended Anthropic negotiations over Dario Amodei's ethical limits
Emails revealed in court documents just revealed why the Pentagon and AI company Anthropic stopped working together.
The big issue? The Pentagon wanted to use Anthropic's AI for any "lawful" purpose, but CEO Dario Amodei pushed back, worried this could include things like domestic surveillance.
When Amodei refused to drop his ethical limits, the defense secretary labeled Anthropic a supply chain risk and ended negotiations.
Disagreement on military AI boundaries
The talks also got stuck because the Pentagon didn't separate offensive from defensive uses for AI, making it tough to draw clear boundaries.
In the end, both sides couldn't agree on how far military AI should go.
This whole situation highlights how tricky (and important) it is to set rules for powerful tech in national security.