OpenAI’s Sam Altman announces Pentagon partnership with ‘security technology’

OpenAI CEO Sam Altman announced late Friday that his company has struck a deal to allow the Department of Defense to use its AI models on the department’s network.

This follows a major rift between the department — also known under the Trump administration as the War Department — and OpenAI competitor Anthropic. Pentagon pushed AI companies, including Anthropic, to allow their models to be used for “all legitimate purposes,” where Anthropic wanted to draw a red line around home monitoring and autonomous devices.

In a lengthy commentary released ThursdayThe CEO of Anthropic, Dario Amodei, said that the company “has not started to oppose other military activities or to try to limit the use of our technology in the military. for this in a way,” but argued that “in the short term, we believe AI could undermine, rather than protect, democratic values.”

More than 60 OpenAI employees and 300 Google employees signed an open letter this week asking employers to support Anthropic.

After Anthropic and the Pentagon failed to reach an agreement, President Donald Trump criticized “Leftwing nut jobs at Anthropic” in social media post which also ordered government agencies to stop using the company’s products after a six-month break.

In separate postSecretary of Defense Pete Hegseth said Anthropic was trying to “usurp veto power over US military decisions.” Hegseth also cited Anthropic as a potential threat: “Effective immediately, no contractor, vendor, or partner that does business with the United States military will do any business with Anthropic.”

On Friday, Anthropic said “It has yet to receive direct communication from the War Department or the White House regarding our negotiations,” but it said it would “defy any threat presented in court.”

Amazing, Altman wrote a post on X that the new OpenAI security agreement includes security to address the same issues that became Anthropic’s focus.

“Our two most important security principles are limiting the number of people in the home and the responsibility of the people in the use of energy, including autonomous devices,” said Altman. “DoW supports these principles, reflects them in laws and policies, and we include them in our agreements.”

Altman said OpenAI will “build security tools to make sure our models are running as they should, which the DoW also wanted,” and will send engineers with the Pentagon “to help with our models and make sure they’re safe.”

“We are asking the DoW to issue the same statement to all AI companies, which in our opinion we think everyone should agree to,” Altman said. “We have shown our strong desire to see the situation escalate away from the law and the authorities and make the necessary agreements.”

Fortune’s Sharon Goldman says that Altman told OpenAI staff at the all-hands meeting that the government would allow the company to create its own “safeguards” to prevent misuse, and that “if the model refuses to work, then the government will not force OpenAI to do the work.”

Altman’s post came shortly before the US and Israeli governments withdrew they will start bombing IranTrump wants to topple the Iranian government.