Facing Backlash, OpenAI Amends Pentagon Deal to Add More Anti-Surveillance Verbiage

OpenAI has amended its controversial Department of Defense contract to include explicit language prohibiting the use of its AI systems for domestic surveillance of U.S. persons. The changes, announced by CEO Sam Altman on Monday, add specific references to the Fourth Amendment, FISA Act, and National Security Act, stating the technology "shall not be intentionally used for domestic surveillance" and prohibits "deliberate tracking, surveillance, or monitoring" through commercial data. This move comes directly after rival Anthropic exited Pentagon negotiations over similar surveillance concerns, with the Times reporting Anthropic demanded a "legally binding promise" not to use its AI on unclassified commercial data that could track locations and browsing histories.

The technical amendments represent a compromise between OpenAI's ethical stance and the Pentagon's position that mass surveillance is already illegal. Pentagon spokesman Sean Parnell argued they merely want "the right to do anything legal" without constraints that could "jeopardize critical military operations." Despite the legal clarification, OpenAI faces significant public backlash including the 'QuitGPT' boycott website claiming 1.5 million participants and celebrity defections like Katy Perry switching to Anthropic's Claude. The situation highlights the growing tension between AI companies' ethical frameworks and government contracts, with OpenAI attempting to navigate both public trust and military partnerships through precise contractual language rather than withdrawing entirely like its competitor.