OpenAI’s Sam Altman announced the company will amend its Defense Department contract to explicitly prohibit domestic surveillance of U.S. citizens. The revised agreement cites the Fourth Amendment, National Security Act, and FISA Act to ban intentional tracking or monitoring of Americans. Altman emphasized the change follows negotiations to clarify restrictions, stating, “Consistent with applicable laws… the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals.”

The move comes amid tensions with Anthropic, which refused Pentagon demands to remove AI guardrails enabling mass surveillance and autonomous weapons. Defense officials had pressured Anthropic to comply, but the company stood firm, calling out “no amount of intimidation or punishment” to alter its ethics stance. Trump’s February 27 order banning federal use of Anthropic’s Claude triggered fallout, with OpenAI finalizing its deal days later.

Post-announcement, Anthropic’s Claude surged to the App Store’s top free app spot, while ChatGPT saw a 295% spike in uninstalls. Altman admitted OpenAI rushed the initial agreement but defended the decision to avoid escalation. He expressed hope the Pentagon would extend similar terms to Anthropic, though details of their contract remain unclear.

The Defense Department’s push for unrestricted AI use contrasts with OpenAI’s ethical positioning, highlighting growing scrutiny over military AI applications. Altman’s memo underscored the importance of transparency in national security partnerships, stressing that rushed agreements risk undermining trust.