OpenAI is revising its recent contract with the US Department of War to include explicit language that prohibits the intentional use of its AI systems for domestic surveillance of U.S. persons, a decision made following backlash over perceived risks associated with the deal. This contract emerged after negotiations with rival startup Anthropic collapsed over the latter’s insistence on strict prohibitions against mass surveillance and lethal autonomous weapons, prompting the Pentagon to pivot to OpenAI. CEO Sam Altman stated that the revisions to the contract aim to clarify these limitations and ensure adherence to existing laws, including the Fourth Amendment, while acknowledging that the rushed rollout of the agreement had appeared “opportunistic and sloppy.”
OpenAI revises Pentagon AI contract to include domestic surveillance ban
