Judge Criticizes HHS for AI-Generated Citations in Grant Changes

A federal judge has sharply criticized the Department of Health and Human Services for citing scientific studies that either do not exist or fail to support claims made in official grant solicitations, raising concerns about the use of artificial intelligence in federal policymaking.

The rebuke centers on sweeping changes HHS made to federal teen pregnancy prevention programs under Secretary Robert F. Kennedy Jr. According to the court, grant solicitations justifying these funding modifications referenced public health studies that appear to be fabricated or misrepresented.

The AI Citation Problem

The judge's written opinion specifically noted that cited research either cannot be located or does not support the conclusions HHS attributed to it. This pattern suggests the agency may have relied on AI-generated references—a growing problem in academic and policy circles where large language models sometimes fabricate plausible-sounding citations to nonexistent papers.

The case highlights a critical vulnerability in government decision-making as agencies increasingly adopt AI tools. When artificial intelligence systems generate false references, they can lend an appearance of scientific legitimacy to policy changes that lack proper evidentiary support.

Why it matters

This ruling exposes how AI tools can undermine the integrity of federal policymaking when used without adequate oversight. Teen pregnancy prevention programs affect millions of young Americans, and funding decisions should rest on verifiable evidence. The incident also sets a precedent for judicial scrutiny of AI-assisted government work, potentially requiring agencies to implement stronger verification protocols before citing research in official documents.

Implications for Federal Agencies

The court's criticism arrives as federal departments face pressure to modernize operations through AI adoption while maintaining rigorous standards for evidence-based policy. HHS now confronts questions about its internal review processes and whether staff verified the existence and relevance of cited studies before publishing grant solicitations.

For organizations applying for federal grants, the ruling underscores the importance of scrutinizing the scientific basis agencies provide for funding criteria changes. Applicants may have grounds to challenge decisions built on phantom research.

The case also illustrates broader risks facing institutions that deploy generative AI without robust fact-checking mechanisms. As these tools become more sophisticated at producing authoritative-sounding text, the potential for inadvertent misinformation in official documents grows.

Details of the judge's rebuke were first reported by The Washington Post. The ruling did not specify what remedies, if any, the court will impose on HHS or whether the agency must revise its grant solicitations with properly verified citations.