OpenAI Backtracks on Pentagon Deal

The controversy ignited just hours after the U.S. government designated rival AI company Anthropic a "supply chain risk." Anthropic had refused a Pentagon deal that didn't include explicit contractual bans on using its AI for mass domestic surveillance or fully autonomous weapons. Public and internal dissent to OpenAI's subsequent agreement was swift and severe. Protests were held outside the company's San Francisco office, and nearly 900 employees from OpenAI and Google signed an open letter demanding their companies refuse military use for surveillance and autonomous killing. The user backlash was measurable, with ChatGPT experiencing a 295% surge in app uninstalls on March 1. Many users reportedly switched to Anthropic's Claude AI system, causing it to climb app store charts. The revised contract now explicitly forbids using OpenAI's systems for the "intentional domestic surveillance of US persons and nationals," a clause that specifically includes a ban on using commercially acquired personal data for tracking people. Furthermore, the amended agreement carves out intelligence agencies. The Pentagon confirmed that organizations like the National Security Agency (NSA) cannot use OpenAI's technology under this deal and would require a separate contract modification to do so. OpenAI had removed a ban on "military and warfare" from its general usage policy in January 2024, a move that foreshadowed its increased willingness to engage with military contracts. In the aftermath, CEO Sam Altman said he was trying to "de-escalate" tensions between the government and AI labs. He also stated that he hoped the Pentagon would offer Anthropic the same amended terms that OpenAI had secured.