OMB Finalizes Guidance on Federal Government AI Procurement

The Office of Management and Budget (OMB) released its final guidance this week on the federal government’s procurement of AI. The guidance comes after President Biden’s Executive Order 14110 on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence tasked OMB with providing initial guidance on the matter, and follows OMB’s previous guidance on federal government use of AI, published in May.

The OMB’s AI procurement memorandum sets out best practices for managing AI risks, including key provisions on interagency information-sharing and collaboration and vendor transparency provisions required during the contracting process. The memo also calls on agencies to promote competition in the AI market and adds additional testing and risk mitigation requirements for procured generative AI systems.

Ensuring the federal government is centering safety and transparency in its procurement of AI is essential because the government is the largest purchaser in the United States economy, spending more than $750 billion each year. As highlighted in EPIC’s 2023 Outsourced & Automated report, much of this funding has gone to contractors providing new, automated tools and services that raise important privacy, equity, and accountability questions. Because of the government’s important role as a purchaser of AI products and services, implementing federal testing and performance standards during the procurement process, as well as baseline transparency and accountability requirements, can improve practices of AI developers across the board.

This week’s guidance follows a March request by OMB for information on responsible government procurement of AI. EPIC and the Brennan Center for Justice led joint comments by a coalition of civil society groups calling on OMB to focus its guidance on creating government procurement processes that protect civil rights and increase transparency.