The General Services Administration (GSA) has unveiled draft terms and conditions aimed at reshaping how artificial intelligence (AI) systems are procured and used by federal agencies. The proposed guidelines, issued by the Federal Acquisition Service (FAS), focus on granting government agencies broader usage rights, ensuring output neutrality, and increasing transparency in AI procurement.
Expanded Government Usage Rights
A key provision in the draft requires vendors providing AI technology to federal agencies to grant the government an irrevocable license to use the systems for any lawful purpose. This measure is intended to prevent contractors from placing restrictions on how agencies can deploy these technologies, giving the federal government the flexibility to utilize AI across various missions and programs.
Neutrality Standards for AI Outputs
The GSA’s proposal also introduces neutrality requirements for AI-generated outputs. Contractors would need to ensure their AI systems produce objective, nonpartisan results when operating in a governmental capacity. According to the draft, this provision is designed to avoid embedding partisan or ideological biases in the outputs produced by AI technologies.
Transparency and Safeguards
Transparency is another focal point of the new guidelines. Under the draft rules, vendors would be required to disclose detailed information about their AI systems, including training methods, limitations, and any modifications made to align with non-U.S. regulations. Additionally, the draft establishes safeguards to protect government data, including restrictions on vendors using federal data for model training without explicit authorization.
Recent Developments in Federal AI Procurement
The GSA’s announcement comes amid growing scrutiny of AI use within the federal government. Recently, the Trump administration mandated the discontinuation of Anthropic’s Claude AI, citing disputes with the Department of War over usage restrictions. Following this, the GSA terminated Anthropic’s OneGov agreement, a decision that Federal Acquisition Service Commissioner Josh Gruenbaum described as ending the company’s availability to all three branches of government through GSA’s pre-negotiated contracts.
Public Feedback and Next Steps
Stakeholders and interested parties have until March 20 to provide comments on the draft terms. The GSA’s efforts reflect a broader movement across federal agencies to enhance oversight and refine procurement rules for AI technologies.
As AI continues to play an increasingly central role in government operations, the new draft terms signal a significant step toward standardizing how these systems are integrated and managed across the public sector.
