WASHINGTON, D.C.: The Trump administration has drafted strict new guidelines for civilian artificial intelligence contracts that would require companies to allow "any lawful" use of their AI systems by the U.S. government, the Financial Times reported.

The proposed rules come amid a dispute between the Pentagon and AI startup Anthropic over restrictions on how its technology can be used.

The Pentagon designated Anthropic a "supply-chain risk" on Thursday and barred government contractors from using the company's AI tools in work related to the U.S. military. The move followed months of disagreement between the Defense Department and Anthropic over safeguards the company wanted placed on its technology.

According to a draft of the guidelines reviewed by the Financial Times, companies seeking U.S. government contracts would have to grant the government an irrevocable license to use their AI models for all lawful purposes.

The guidance, being developed by the General Services Administration (GSA), would apply to civilian contracts and forms part of a wider government effort to tighten how federal agencies procure AI services, the newspaper said. The rules are also expected to mirror measures the Pentagon is considering for military contracts.

"It would be irresponsible to the American people and dangerous to our nation for GSA to maintain a business relationship with Anthropic," said Josh Gruenbaum, commissioner of the Federal Acquisition Service. This GSA unit helps procure software for federal agencies.

"As directed by the President, GSA has terminated Anthropic's OneGov deal - ending their availability to the Executive, Legislative, and Judicial branches through GSA's pre-negotiated contracts," Gruenbaum said.

The White House did not immediately respond to requests for comment.

The draft guidelines also include provisions governing how AI systems should be designed and disclosed. According to the Financial Times report, contractors would be required to ensure that they "must not intentionally encode partisan or ideological judgments into the AI systems' data outputs."

Companies would also need to disclose whether their AI models have been "modified or configured to comply with any non-U.S. federal government or commercial compliance or regulatory framework," the newspaper said.