(Credit: CFOTO/Future Publishing via Getty Images)
The Pentagon is considering ending its work with Anthropic over the AI company's insistence that some of its Claude safeguards remain in place for military operations.
As The Wall Street Journal reports, the Defense Department used Claude (via its Palantir contract) to help it carry out the attack on Venezuela and capture former President Nicolás Maduro. An Anthropic employee then asked Palantir for details on what happened, the WSJ says, though a spokesperson insists the company only has "routine discussions on strictly technical matters...with industry partners, including Palantir."
"Anthropic's conversations with the [Defense Department] to date have focused on a specific set of Usage Policy questions—namely, our hard limits around fully autonomous weapons and mass domestic surveillance—none of which relate to current operations," a spokesperson tells Axios.
In 2024, Anthropic signed a deal with Palantir to have Claude "support government operations," including data processing, identifying trends, and "helping US officials to make more informed decisions in time-sensitive situations." Last year, Anthropic signed on for Palantir’s FedStart tool, which allows Anthropic to offer Claude to federal government employees, who can use it to "enhance their efficiency in writing, analyzing data, and solving complex problems."
Anthropic also has a $200 million contract with the Pentagon, but its concerns about how the government uses Claude have reportedly put that contract at risk.
Claude is accessible on the Pentagon's classified networks, Axios says, and the Defense Department is negotiating to also allow OpenAI's ChatGPT, Google's Gemini, and xAI's Grok. They would have to agree to let the military use their AI for "all lawful purposes," but Anthropic is reportedly insisting that Claude not be used for "the mass surveillance of Americans and fully autonomous weaponry."
“Our nation requires that our partners be willing to help our warfighters win in any fight,” a Pentagon spokesperson told the WSJ.
"Everything's on the table," including the cancellation of the Anthropic contract, a senior administration official told Axios. "But there'll have to be an orderly replacement [for] them, if we think that's the right answer."