Sam Altman admits OpenAI looked “opportunistic and sloppy.” Those are his words, not his critics'.

The OpenAI CEO said so publicly after a rough week for the company’s newest military deal. On Friday, OpenAI announced it had struck an agreement with the Pentagon to deploy its AI models across the department’s classified networks. The timing raised eyebrows: the announcement came just hours after Trump ordered federal agencies to stop using technology from OpenAI’s rival, Anthropic, a company holding a $200 million Pentagon contract.

It wasn’t just the timing that drew fire. Under the original deal, OpenAI agreed to let the Pentagon use its AI systems for any lawful purpose. Altman had touted two safety principles baked into the agreement: prohibitions on domestic mass surveillance, and human responsibility for the use of force, including autonomous weapons. “The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement,” he wrote on X on Friday. Critics weren’t convinced the language had teeth.

What the amended OpenAI Pentagon contract says

After a weekend of criticism, OpenAI went back and rewrote the contract. The amended deal now states, in plain language, that the AI system “shall not be intentionally used for domestic surveillance of U.S. persons and nationals,” consistent with applicable laws including the Fourth Amendment.

The contract goes further. “For the avoidance of doubt, the Department understands this limitation to prohibit deliberate tracking, surveillance, or monitoring of US persons or nationals, including through the procurement or use of commercially acquired personal or identifiable information,” the amended text reads, according to Altman’s post.

OpenAI also confirmed a new carve-out for the spy agencies: its services will not be used by Department of War intelligence agencies, including the National Security Agency. Any such use would require a follow-on change to the contract.

“It’s critical to protect the civil liberties of Americans, and there was so much focus on this, that we wanted to make this point especially clear,” Altman wrote on social media. He added, in classic iterate-in-public fashion: “Just like everything we do with iterative deployment, we will continue to learn and refine as we go.”

“Opportunistic and sloppy”: Altman’s mea culpa

Altman didn’t just fix the contract. He apologized for how the whole thing landed. Writing on X, he said the company should not have hurried to announce the agreement on 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 company will hold an all-hands meeting on Wednesday morning to answer more questions from employees, some of whom are clearly not thrilled. According to reporting by AI Base, more than 60 OpenAI employees signed a joint letter this week backing Anthropic’s hard line on military use, the very stance the Pentagon punished.

That’s a remarkable internal split. OpenAI’s own staff are siding with the company the Trump administration just tried to freeze out of the defense supply chain.

Why Anthropic is at the center of this

This story doesn’t make sense without the Anthropic feud. The Pentagon had spent months pressing Anthropic to strip safety guardrails from Claude and grant the military unfettered access to its AI products. Anthropic refused, citing concerns about mass domestic surveillance and autonomous weapons.

The administration’s response was forceful. Trump ordered federal agencies to halt their use of Anthropic’s technology, and the Pentagon designated the company a supply chain risk, effectively urging defense contractors to cut ties with it. Anthropic called the move “legally unsound” and a “dangerous precedent,” and says it will challenge the designation in court.

A Pentagon statement struck a very different tone about the two companies: “We will always come to the table for reasonable discussion as we did with OpenAI. Anthropic didn’t want to do that, because they have their own personal vendettas.”

Here’s the twist. Altman’s safety principles are essentially the same red lines Anthropic drew: no mass surveillance, humans accountable for lethal force. Even Altman sees the irony. Over the weekend, he repeated his objection to the Pentagon’s punishment of Anthropic: “I reiterated that Anthropic should not be designated as a SCR, and that we hope the DoW offers them the same terms we’ve agreed to,” he said.

Why this matters

Three things are colliding here, and none of them are going away.

First, the Pentagon has now made explicit what it wants from AI labs: broad access, with the lab’s own guardrails negotiable. Anthropic said no and got blacklisted. OpenAI said yes and got a contract. Every other frontier lab (Google, xAI, Meta) is watching that incentive structure and doing the math.

Second, OpenAI’s fix shows the pressure points actually work. A weekend of backlash produced the clearest ban on domestic surveillance any AI company has extracted from a military contract: Fourth Amendment, named federal laws, no NSA use. Vague “principles” got converted into contract language because people noticed the gap.

Third, this is now a litigation story. Anthropic plans to challenge its supply-chain-risk designation in court, and the amended OpenAI contract may become evidence. If the Pentagon offered OpenAI terms it denied Anthropic, a judge is going to want to know why.

FAQ

Did OpenAI change its Pentagon contract?

Yes. After a weekend of criticism, OpenAI amended the agreement to state its AI systems “shall not be intentionally used for domestic surveillance of U.S. persons and nationals,” consistent with the Fourth Amendment and federal law.

Will OpenAI’s AI be used for domestic surveillance?

No. The amended contract bans deliberate tracking, surveillance, or monitoring of US persons, including through commercially acquired personal data, and bars use by intelligence agencies like the NSA without a follow-on contract change.

Why did Sam Altman call the deal “opportunistic and sloppy”?

Altman admitted OpenAI rushed to announce the Pentagon deal on Friday, hours after President Trump blacklisted rival Anthropic, making the timing look self-serving. He said the company was “genuinely trying to de-escalate things.”

Is Anthropic still fighting the Pentagon?

Yes. Anthropic called the supply-chain-risk designation “legally unsound” and a “dangerous precedent” and plans to challenge it in court. Altman has publicly said Anthropic should not carry that designation.

Sources: East Bay Times (Tribune wire), The Economic Times, CNBC-TV18, AI Base, The Hindu Business Line, Bloomberg News.