Undisclosed AI Use in Federal Bid Evaluations Heightens Protest Risks
Federal agencies’ undisclosed use of AI in contract evaluations raises legal protest risks.
Why it matters: Legal and compliance teams advising government contractors must address AI’s impact on transparency and fairness in procurement. Understanding these risks helps manage potential bid protests and sanctions in an evolving regulatory landscape.
- Federal agencies use generative AI to assist contract proposal evaluation but often do not disclose this use in solicitations.
- AI-produced analysis risks include factual errors, inconsistent judgments, and biased or non-independent decisions.
- The Government Accountability Office (GAO) warns that improper use of AI in bid protests can result in dismissal and sanctions.
- The Federal Circuit’s en banc ruling in Percipient.ai clarified that only actual or prospective bidders have standing to file bid protests under the Tucker Act.
Federal agencies increasingly employ generative artificial intelligence (AI) tools to evaluate contract proposals by summarizing submissions, identifying strengths and weaknesses, assessing compliance, drafting technical findings, and supporting source selection decisions. However, agencies often do not disclose AI’s involvement in solicitations, raising concerns about transparency.
AI-generated outputs may contain factual inaccuracies, utilize undisclosed evaluation criteria, yield inconsistent assessments, and lead to biased or non-independent decisions. Contractors may spot these issues during debriefings through generic or repetitive language, oversimplified summaries, references to irrelevant concepts, or conclusions that diverge from solicitation requirements.
Aron C. Beezley, a federal procurement attorney, explains that "when agencies rely on undisclosed or poorly governed AI tools during federal procurement evaluations, contractors may have grounds to challenge the procurement at the Government Accountability Office (GAO) or the U.S. Court of Federal Claims (COFC)." This introduces new grounds for bid protests related to AI’s impact on fairness and transparency.
The GAO has issued warnings about misuse of generative AI in protests. As David P.J. Timm notes, "Protesters beware – the Government Accountability Office (GAO) wants you to know they can and will sanction misuse of Generative Artificial Intelligence (GenAI)." This highlights the importance of carefully managing AI’s role in legal challenges to avoid dismissal or sanctions.
Clarity on who can bring bid protests has also evolved. The Federal Circuit’s en banc decision in Percipient.ai, Inc. v. United States restricted protest standing to actual or prospective bidders, consistent with the Tucker Act’s requirements governing the Court of Federal Claims. Dawn Stern observes that this ruling reversed prior broader interpretations, narrowing the pool of eligible protesters.
Together, these developments underscore the need for contractors and legal teams to stay vigilant about undisclosed AI use in procurement and understand the procedural requirements governing bid protests. Demanding transparency about AI’s role in evaluations is vital to managing both substantive challenges and compliance risks.
By the numbers:
- 2023 — GAO issued guidance warning against misuse of generative AI in bid protests
- March 17, 2023 — Date of the Federal Circuit en banc ruling in Percipient.ai, Inc. v. United States
- Only actual or prospective bidders — standing requirement clarified by the Federal Circuit for bid protests
Yes, but: While AI use presents new legal risks, the government also has strong incentives to maintain fairness and transparency, and improper AI use can lead to sanctions and protest dismissals, which may deter misuse.
What's next: Legal experts expect further guidance from federal agencies on AI disclosure policies in procurement and possible legislative updates addressing AI use in government contracting.