OpenAI Amends Pentagon Contract Following Employee and Public Backlash Over Surveillance Concerns

Summary

OpenAI has announced it is amending its contract with the Pentagon following significant public and employee backlash over concerns that the company's AI systems could be used for mass domestic surveillance or lethal autonomous weapons. CEO Sam Altman posted an internal memo stating that the revised agreement will explicitly prohibit the use of OpenAI's AI systems for domestic surveillance of US persons, consistent with the Fourth Amendment, and will exclude use by Department of War intelligence agencies like the NSA without additional contract modifications.

The controversy erupted after OpenAI struck a deal with the Pentagon on Friday to deploy its AI models on classified military networks, coming just one day before a US strike on Iran. The timing and nature of the deal drew sharp criticism, with nearly 500 OpenAI and Google employees signing an open letter supporting Anthropic's decision to reject a similar Pentagon contract. Anthropic had walked away from negotiations after the Pentagon refused to include explicit contractual bans on mass domestic surveillance and fully autonomous weapons systems.

In his memo, Altman acknowledged that OpenAI "got things wrong" and should not have "rushed" to seal the deal, admitting it appeared "opportunistic and sloppy." The OpenAI agreement came immediately after President Trump ordered federal agencies to halt use of Anthropic's Claude system following the breakdown in talks. Protests have erupted at OpenAI offices in San Francisco and London, with advocacy group QuitGPT launching a boycott and organizing further demonstrations. The incident highlights growing tensions within the AI industry over military applications and the ethical boundaries companies are willing to set.

• Protests at OpenAI offices and a boycott campaign by QuitGPT reflect broader concerns about AI companies' military partnerships and ethical boundaries