Washington D.C. Partner David Timm authored an article for the National Contract Management Association analyzing the growing misuse of generative artificial intelligence in federal procurement disputes and the risks it poses to tribunals, agencies, and taxpayers. The article explores how large language models are increasingly used to generate error-filled bid protests and appeals, including filings that contain fabricated facts and citations.

David highlights the case of Oready, LLC as a cautionary example, detailing how the contractor filed a barrage of pro se protests and appeals that ultimately prompted sanctions from the Government Accountability Office (GAO). He explains that while some GenAI misuse is unintentional, a more concerning trend involves contractors deliberately exploiting LLMs to file essentially cost-free, frivolous protests that delay procurements and consume significant public and private resources.

The article also examines how procurement tribunals, including GAO, the Court of Federal Claims, and boards of contract appeals, have begun to address GenAI misuse, often focusing narrowly on hallucinated legal citations while overlooking broader factual inaccuracies. David outlines the real-world consequences of these filings, including prolonged protest timelines, stayed procurements, and increased burdens on agencies, awardees, and the courts.

To mitigate these risks, David proposes specific reforms, including disclosure and certification requirements for filings prepared with GenAI tools, enhanced sanctions, and potential anti-filing injunctions for repeat offenders. His analysis underscores the need for procurement tribunals to adapt quickly to emerging technology risks to protect the integrity and efficiency of the federal procurement process.