Trump Admin Drafts 'Any Lawful Use' AI Rules for Agencies

New Rules Would Reshape How Government Buys AI

The Trump administration is drafting sweeping new guidelines for civilian artificial intelligence contracts that would fundamentally change the relationship between AI companies and the federal government. At the core of the proposed rules: companies pursuing federal contracts must grant the U.S. government an irrevocable license to use their AI systems for all lawful purposes, with no ability to restrict specific applications after the sale.

The guidance, being developed by the General Services Administration (GSA), would apply initially to civilian agency contracts and forms part of a wider government effort to tighten how federal agencies procure AI services. Similar requirements are being considered for military procurement contracts, which would extend the framework across the entire federal government.

Political Neutrality and Foreign Compliance Disclosure

Beyond the broad licensing requirement, the draft guidelines introduce significant new mandates around AI system behavior. Contractors would be required to ensure their models deliver "a neutral, non-partisan tool that does not manipulate responses in favour of ideological dogmas such as diversity, equity, inclusion." The language explicitly targets what the administration views as built-in political bias in current AI systems.

Companies would also need to disclose whether their models have been modified to comply with foreign regulatory standards, such as the European Union's Digital Services Act or AI Act. This provision appears designed to prevent AI companies from applying international content moderation standards to U.S. government systems, ensuring that models deployed for federal use reflect American policy priorities rather than foreign regulatory frameworks.

The Anthropic Dispute as Catalyst

The proposed rules are a direct response to the Pentagon's confrontation with Anthropic, which refused to grant unrestricted access to its Claude AI for military applications. Defense Secretary Pete Hegseth demanded Anthropic allow "unrestricted use" for "all lawful purposes." When Anthropic insisted on carve-outs for mass domestic surveillance and fully autonomous weapons, the administration designated the company a "supply chain risk" and terminated a contract worth over $200 million.

GSA Commissioner Josh Gruenbaum defended the aggressive stance, stating: "It would be irresponsible to the American people and dangerous to our nation for GSA to maintain a business relationship with Anthropic." The GSA also terminated Anthropic's participation in the OneGov program, which had allowed federal institutions to access its technology through pre-negotiated contracts.

The contrast with OpenAI's approach is instructive. OpenAI's Pentagon contract permits the Department of Defense to use the AI system for "all lawful purposes, consistent with applicable law." However, OpenAI retains "full discretion" over safety classifiers and maintains practical restrictions through cloud-only deployment, cleared personnel involvement in monitoring, and termination rights if the government violates agreement terms. The distinction between Anthropic's contractual restrictions and OpenAI's operational restrictions has become a central debate in AI governance.

Legal Scholars Raise Governance Concerns

Not everyone views the proposed rules as a straightforward improvement. Jessica Tillipman, a government contracts law professor at George Washington University, argues that the real problem is not whether companies should restrict government use but that commercial acquisition pathways systematically limit the government's ability to negotiate protective terms it actually needs, including transparency requirements, audit rights, data protections, and safeguards against contractor lock-in.

Tillipman warns that designating Anthropic a supply chain risk for aggressive contract negotiation may deter future companies from pushing back on inadequate government terms, ultimately worsening governance outcomes despite faster procurement. The precedent could lead to a chilling effect where AI companies simply accept whatever terms the government demands rather than risk blacklisting.

What This Means for the AI Industry

If finalized, the "any lawful use" framework would force every AI company seeking federal business to make a fundamental choice: accept unrestricted government access as a condition of doing business, or walk away from hundreds of millions in federal contracts. For AI engineers and professionals, the rules could reshape which companies pursue government work and how safety teams operate within those organizations. The Commerce Department's related report on state AI laws, due March 11, is expected to further clarify the administration's vision for a unified national AI policy framework.