OpenAI Details Its AI Agreement with the Pentagon
Following the US government's ban on rival Anthropic, OpenAI has shared more details about its deal to provide the Pentagon with access to its advanced AI models. CEO Sam Altman emphasized transparency and oversight in the collaboration, which marks a significant milestone in the use of commercial AI for national security applications.
OpenAI's agreement with the Pentagon includes specific "red lines" that prohibit the use of its AI for mass domestic surveillance, directing autonomous weapons, or making high-stakes automated decisions. These are reportedly the same safeguards that rival AI company Anthropic insisted on before the U.S. government terminated its contract. The deal materialized just hours after the Trump administration ordered all federal agencies to stop using technology from Anthropic, designating the company a "supply chain risk to national security." This move followed a public standoff where Anthropic refused to remove its safeguards for a $200 million Defense Department contract. OpenAI's models will be deployed in a cloud-only environment, meaning the AI will remain on servers controlled by the company, not embedded directly into military hardware. This architecture, along with the involvement of cleared OpenAI personnel, is intended to ensure compliance with the agreed-upon safety restrictions. The arrangement is part of a broader "OpenAI for Government" initiative, which consolidates the company's collaborations with various U.S. government entities, including NASA and the Air Force Research Laboratory. In a public Q&A session, OpenAI CEO Sam Altman admitted the Pentagon deal was "definitely rushed" and that the "optics don't look good." He stated the agreement was an attempt to de-escalate the situation between the tech industry and the Department of Defense. Altman also expressed concern over the precedent set by the government's blacklisting of Anthropic, calling it "an extremely scary precedent." He has publicly asked the Pentagon to offer the same contractual terms to all AI companies to help resolve the situation with Anthropic.
