Altman Admits Pentagon Deal Looked 'Sloppy'
OpenAI CEO Sam Altman has admitted the company's recent Pentagon deal "looked opportunistic and sloppy." The comment addresses the perception created by how quickly OpenAI secured the contract after its rival, Anthropic, was blacklisted by the government. Altman's candor highlights the unusual speed and opacity of the deal, which has drawn scrutiny over the consolidation of AI suppliers to the U.S. government.
The blacklisting of Anthropic, a major AI provider for classified military systems, occurred after the company refused to remove two contractual guardrails: a ban on mass domestic surveillance of U.S. citizens and a prohibition on fully autonomous weapons without human oversight. Defense Secretary Pete Hegseth then designated the firm a "supply chain risk to national security," a label previously reserved for foreign entities like Huawei. Hours after Anthropic was blacklisted, OpenAI announced its own deal with the Pentagon. The Trump administration had ordered all federal agencies to cease using Anthropic's technology, with President Trump labeling the company's executives "leftwing nut jobs" for their stance. Anthropic's now-terminated contract with the Department of Defense was valued at up to $200 million. The company has vowed to challenge its "supply chain risk" designation in court, calling the move legally unsound and an unprecedented action against an American company. In response to public backlash and a "delete ChatGPT" campaign online, OpenAI amended its new agreement. The updated terms explicitly bar the use of its technology for mass domestic surveillance and confirm that intelligence agencies like the NSA would require a separate agreement to use their services. The controversy led to a surge in downloads for Claude, Anthropic's chatbot, which briefly surpassed ChatGPT to become the top app on Apple's App Store. Meanwhile, hundreds of employees from major tech companies, including Google and Microsoft, signed an open letter urging their employers to refuse the Pentagon's demands. OpenAI has stated its agreement contains more safety guardrails than Anthropic's did and that it retains "full control" over its safety mechanisms. Sam Altman argued that Anthropic may have sought more operational control than OpenAI, which felt comfortable relying on existing laws as a framework for its "red lines" on surveillance and autonomous weapons. Legal experts, however, have raised doubts about how enforceable OpenAI's safeguards are, suggesting the contractual language offers little limitation beyond "all lawful use." They note that existing policies referenced in the contract can be modified and that there isn't a precise legal definition of "mass domestic surveillance."
