EPIC, Coalition Urge OMB to Adopt AI Procurement Guidelines with Civil Rights and Transparency in Mind

Today, a cross-cutting coalition of civil society organizations led by EPIC and the Brennan Center for Justice filed joint comments to the White House Office of Management and Budget (OMB) laying out the need for additional federal AI procurement guidance to adequately mitigate the risks of procured AI. The comments use real-world examples of harmful AI contracts, such as U.S. Customs and Border Protection (CBP)’s social media monitoring contract, to operationalize key components of the OMB’s recently published federal AI risk management guidance.

Even though agencies regularly procure and use private-sector AI systems, standard procurement and contract management practices fail to address the unique risks and informational needs that agencies face when adopting AI systems. For example, the Federal Acquisition Regulation (FAR)’s requirements do not provide agencies with sufficient information to monitor AI contracts, nor do they provide agencies with the flexibility they need to adapt to new AI risks during the contract term. To address AI risks and align procurement practices with the OMB’s AI risk management guidance, the EPIC- and Brennan Center-led coalition called on OMB to require:

1. A pre-solicitation needs assessment to identify anticipated AI risks and evaluate whether non-AI alternatives would better meet agency needs;
   
2. Vendor disclosures of all information necessary to comply with the OMB’s guidance on federal AI risk management; and

3. An AI risk register, which agencies can use to track and incorporate vendor- and system-specific risks into contract language.

This coalition comment is the latest in a series of EPIC efforts to improve the ways government acquires and uses private-sector AI systems, including EPIC’s previous OMB comments, our 2023 investigative report into state AI procurement trends and recent advocacy around California’s AI procurement guidelines.