Pentagon moves away from Anthropic Claude, testing OpenAI and Google
The Pentagon is moving away from using Anthropic's Claude AI in its secret military projects, after concerns were raised about supply chain risks tied to Anthropic's strict safety rules.
Now, they're testing out alternatives from OpenAI and Google, aiming to switch over in the next six months, even though Anthropic is fighting the decision in court.
Pentagon tests AIs with 25 users
To find the best fit, 25 top users across five military commands are putting these new AIs to the test.
The Pentagon says using multiple systems will help avoid over-reliance on any one company.
While some human rights groups worry about how fast AI is moving into warfare, officials believe smarter tech will boost efficiency for routine tasks, and they're sticking with their plan to move forward.