US defense calls Anthropic a 'security risk' over AI talks
Talks between Anthropic and the US Defense Department broke down on Friday, after Anthropic refused to give the Pentagon full access to its AI systems—citing concerns about possible surveillance of Americans.
In response, Defense Secretary Pete Hegseth labeled Anthropic a "security risk," while the article notes the distinct "supply chain risk" designation had never been used against an American company.
OpenAI steps in as Anthropic plans to fight label
With Anthropic out, OpenAI quickly stepped in and reached an agreement with the Pentagon that satisfied their rules for lawful and safe AI use.
While OpenAI got the deal, it's also facing some ethical questions about working with the government.
Meanwhile, Anthropic plans to fight its new label in court—highlighting just how complicated using AI in national security has become.