OpenAI Amends Pentagon Deal After Backlash
OpenAI is backtracking on its landmark deal with the U.S. Department of Defense after intense public backlash. CEO Sam Altman admitted the original agreement looked "sloppy" and is now adding language to strengthen anti-surveillance safeguards. Altman specifically said "two sentences" would be added to clarify the ethical use of its AI in military contexts.
The controversial deal materialized shortly after rival AI company Anthropic refused a Pentagon contract. Anthropic specifically objected to demands that would have removed safeguards against using its AI for mass domestic surveillance and fully autonomous weapons, leading the Trump administration to blacklist the company as a "supply chain risk." This partnership followed a key change in OpenAI's own policies. In January 2024, the company quietly removed language from its usage guidelines that broadly banned "military and warfare" applications, replacing it with a more specific prohibition against using the technology to "develop or use weapons." The backlash was swift from both inside and outside the company. Nearly 100 OpenAI employees, along with hundreds from Google, signed an open letter calling on their employers to refuse demands for AI-powered surveillance and autonomous killing. Publicly, a "delete ChatGPT" campaign gained traction on social media, while Anthropic's chatbot, Claude, surged to the top of Apple's App Store charts. In response, OpenAI added specific language to the agreement explicitly prohibiting the intentional use of its AI for domestic surveillance of U.S. persons, referencing the Fourth Amendment and other laws. The amended contract also bars use by intelligence agencies like the NSA without a follow-on modification to the agreement. Despite the changes, critics noted the new language only forbids the "intentional" surveillance of Americans, which could leave loopholes for "incidental collection" of data. During an internal all-hands meeting, Sam Altman reportedly told staff that while OpenAI can build the technology responsibly, the company ultimately cannot control how the U.S. military decides to use it once it is deployed.
