OpenAI Revises Defense Deal Amid Surveillance Concerns, Public Backlash
WASHINGTON – OpenAI will amend its contract with the U.S. Department of Defense to explicitly prohibit the use of its artificial intelligence tools for domestic mass surveillance, CEO Sam Altman announced Monday, following a wave of criticism from the public and its own employees. The move comes after the company quickly stepped in to fill a gap left by Anthropic, which was dropped from the Pentagon contract after refusing similar assurances.
The initial deal, announced Friday, drew immediate scrutiny, with critics questioning how OpenAI secured ethically acceptable terms when Anthropic had not. Altman acknowledged the rollout was rushed and “looked opportunistic and sloppy,” according to an internal message to staff.
“We shouldn’t have rushed to get this out on Friday. The issues are super complex, and demand clear communication,” Altman said.
The revised contract includes specific language stating the AI system “shall not be intentionally used for domestic surveillance of U.S. persons and nationals.” The Pentagon also understands this prohibits the deliberate tracking or monitoring of American citizens, including through commercially acquired data, according to Altman’s statement.
The controversy ignited a broader debate about the ethical implications of AI in defense and sparked an open letter signed by over 900 employees of OpenAI and Google, urging both companies to resist pressure to allow their models to be used for domestic surveillance or autonomous weapons systems.
“They’re trying to divide each company with fear that the other will give in,” the letter stated. “This letter serves to create shared understanding and solidarity in the face of this pressure.”
The backlash was reflected in user data, with a surge in uninstalls of OpenAI’s ChatGPT app and a corresponding increase in downloads of Anthropic’s Claude app, which briefly became the top app on the U.S. App Store.
The Pentagon’s partnership with Anthropic soured in 2025 when the company sought guarantees its technology wouldn’t be used against U.S. citizens or for autonomous weapons. Following Anthropic’s refusal to comply, Defense Secretary Pete Hegseth labeled the company a supply-chain threat, and President Donald Trump subsequently banned federal agencies from using Anthropic’s AI solutions, with a six-month phase-out period for existing contracts. Trump warned of “major civil and criminal consequences” should Anthropic not cooperate.
Altman insisted the Pentagon had agreed that agencies like the National Security Agency would not utilize OpenAI’s tools, citing the technology’s current limitations and the need for a deeper understanding of safety tradeoffs. He added that OpenAI and the Pentagon would collaborate on developing technical safeguards.
Former OpenAI policy research head Miles Brundage expressed skepticism, suggesting OpenAI may have compromised ethical principles while portraying a more favorable outcome. “OpenAI employees’ default assumption here should unfortunately be that OpenAI caved + framed it as not caving, and screwed Anthropic while framing it as helping them,” he said.
