CEO Sam Altman acknowledges that OpenAI appeared "opportunistic and sloppy" in a hurried agreement with the Pentagon

OpenAI chief executive Sam Altman has admitted that the company appeared “opportunistic and sloppy” in rushing to secure a Pentagon contract for the use of its advanced AI systems in classified environments.

Posting on X, Altman said OpenAI should not have hurried to announce the agreement last Friday. “We were genuinely trying to de-escalate things and avoid a much worse outcome, but I think it just looked opportunistic and sloppy. Good learning experience for me as we face higher-stakes decisions in the future,” he wrote.

The deal was unveiled only hours after rival AI firm Anthropic refused to allow unconditional military use of its Claude models, drawing criticism from US officials. In a separate post, Altman reiterated his objection to the Pentagon’s decision to designate Anthropic a “supply chain risk,” a label that effectively pressures defence contractors to avoid working with the company. He said he had urged the Department of War to offer Anthropic the same terms agreed with OpenAI.

Altman also announced that an all-hands meeting would be held on Wednesday to address employee concerns.

Amid mounting scrutiny, Altman confirmed that OpenAI would revise its contract to clearly state that its AI systems will not be used for domestic surveillance. Citing the Fourth Amendment, he wrote that the AI system “shall not be intentionally used for domestic surveillance of US persons and nationals.” He added that this limitation prohibits deliberate tracking, surveillance or monitoring of Americans, including through commercially acquired personal data.

He further confirmed that OpenAI’s services would not be used by intelligence agencies such as the National Security Agency unless a follow-on amendment to the contract is made.

The dispute began when the Pentagon sought access to Anthropic’s AI models for sensitive uses, including weapons development and intelligence operations. Anthropic chief executive Dario Amodei resisted, warning that such an arrangement could enable mass domestic surveillance or autonomous weapons, potentially jeopardising the company’s $200 million July 2025 contract with the Department of War.

After Anthropic declined, the Pentagon under US President Donald Trump designated it a supply chain risk. The following day, OpenAI announced its own agreement, granting the department access to its AI systems for all lawful uses.

Altman later said during an Ask Me Anything session that entering the agreement was a calculated risk. If it helps reduce friction and stabilise relations between the government and the AI industry, he said, it would justify the reputational cost. If not, OpenAI would have to accept criticism for moving too quickly.