The Pentagon is considering terminating a partnership with artificial intelligence company Anthropic over its insistence on setting some limits on how the military can use its models, according to a senior U.S. government official. The Pentagon is urging four leading artificial intelligence labs to allow the military to use their tools for "all lawful uses," including even the most sensitive areas such as weapons development, intelligence collection and battlefield operations.
Anthropic has yet to agree to those terms, and after months of difficult negotiations, the Pentagon is growing impatient.
Anthropic reportedly insists that two areas remain off-limits: mass surveillance of Americans and fully autonomous weapons.
Officials familiar with the matter revealed that "anything is possible," including reducing the scale of cooperation with Anthropic or even terminating the cooperation entirely.
An Anthropic spokesperson responded that the company remains "committed to leveraging cutting-edge artificial intelligence technology to support U.S. national security."