OpenAI Announces Pentagon AI Deal with Enhanced Safety Measures

Artificial intelligence company OpenAI revealed details about its new Pentagon agreement, emphasizing stronger safety protections than previous military AI contracts. The deal comes after the Trump administration cut ties with rival AI company Anthropic over technology safety concerns.

Artificial intelligence company OpenAI disclosed on Saturday that its newly announced Pentagon partnership includes enhanced protective measures for military applications on classified government networks.

The announcement follows President Trump’s Friday directive ordering the government to cease collaboration with AI competitor Anthropic. The Pentagon subsequently declared it would designate the startup as a supply-chain threat, delivering a significant setback to the artificial intelligence company following disputes over technology safety protocols. Anthropic has indicated it will legally contest any risk classification.

Microsoft, Amazon, and SoftBank-backed OpenAI revealed its Pentagon arrangement late Friday, shortly after the Anthropic developments.

“We think our agreement has more guardrails than any previous agreement for classified AI deployments, including Anthropic’s,” OpenAI stated on Saturday.

The artificial intelligence company outlined that its contract with the Defense Department – which the Trump administration has rebranded as the Department of War – establishes three prohibited uses: OpenAI’s technology cannot support widespread domestic surveillance operations, control autonomous weapons systems, or handle critical automated decision-making processes.

“In our agreement, we protect our red lines through a more expansive, multi-layered approach. We retain full discretion over our safety stack, we deploy via cloud, cleared OpenAI personnel are in the loop, and we have strong contractual protections,” the company explained.

Over the past year, the Pentagon has established agreements valued at up to $200 million each with leading AI companies, including Anthropic, OpenAI and Google. Military officials aim to maintain maximum operational flexibility for defense purposes without being constrained by technology companies’ concerns about unreliable AI powering weapons systems.

OpenAI warned that any contract violations by the federal government could result in agreement termination, though the company added, “We don’t expect that to happen.”

The company also defended competitor Anthropic against the supply-chain risk designation, stating, “We have made our position on this clear to the government.”