OpenAI shares more details about its agreement with the Pentagon | TechCrunch

According to CEO Sam Altman himself, OpenAI’s deal with the Department of Defense was “absolutely rushed” and “the optics don’t look good.”

After negotiations between Anthropic and the Pentagon collapsed on Friday, President Donald Trump ordered federal agencies to stop using Anthropic’s technology. a transition period of six monthsand Defense Secretary Pete Hegseth said he was labeling the AI ​​company as a supply chain risk.

OpenAI then quickly announced that it had struck its own deal on deploying models in classified environments. With Anthropic saying it was drawing red lines around the use of its technology in fully autonomous weapons or mass domestic surveillance, and Altman saying OpenAI had the same red lines, there were some obvious questions: Was OpenAI honest about its safeguards? Why was it able to reach a deal while Anthropic did not?

So while OpenAI executives defended the deal on social media, the company also published a blog post explaining the approach.

In fact, the post pointed to three areas where OpenAI’s models cannot be used: mass domestic surveillance, autonomous weapons systems, and “high-stakes automated decisions (e.g. systems like ‘social credit’).”

The company said that unlike other AI companies that have “reduced or removed their security rails and relied primarily on usage policies as their primary safeguards in national security deployments,” OpenAI’s agreement protects its red lines “through a more comprehensive, multi-layered approach.”

“We maintain full discretion over our security stack, we deploy via the cloud, approved OpenAI staff are informed, and we have strong contractual protections,” the blog said. “This is all in addition to the strong existing protections in U.S. law.”

The company added: “We don’t know why Anthropic has not been able to reach this deal, and we hope they and more labs will consider doing so.”

After the message was published, Techdirt’s Mike Masnick claimed that the deal “allows for absolute domestic surveillance,” because it says the collection of private data will comply with this Executive Order 12333 (along with a number of other laws). Masnick described that order as “how the NSA conceals its domestic surveillance by capturing communications by tapping lines *outside the US*, even if it contains information from/about US persons.”

In a LinkedIn messageOpenAI’s head of national security partnerships Katrina Mulligan argued that much of the discussion surrounding the contract language assumes that “the only thing standing between Americans and the use of AI for mass domestic surveillance and autonomous weapons is a single use policy provision in a single contract with the Department of War.”

“That’s not how any of this works,” Mulligan said, adding: “Deployment architecture is more important than contract language […] By limiting our efforts to cloud API, we can ensure that our models cannot be directly integrated into weapon systems, sensors or other operational hardware.”

Altman also asked about the deal on X, where he admitted it was a rush and resulted in significant backlash against OpenAI (so much so that Anthropic’s Claude overtook OpenAI’s ChatGPT in Apple’s App Store on Saturday). So why would you do it?

“We really wanted to de-escalate things and we thought the deal offered was good,” Altman said. “If we are right and this does indeed lead to a de-escalation between the [Department of War] and the industry, we will look like geniuses, and a company that has put a lot of effort into doing things to help the industry. If not, we will continue to be characterized as […] hurried and inattentive.”