Court Bars HHS from Using AI-Generated False Citations in Grants
A federal judge issued a preliminary injunction halting HHS's AI-generated false citations in grant solicitations.
Why it matters: Federal agencies increasingly use AI in policymaking and funding decisions, raising risks of misinformation. Legal professionals should watch how courts address AI accountability and accuracy in government documents.
- On Sept. 1, 2026, U.S. District Judge Christopher Cooper issued a preliminary injunction against HHS's Teen Pregnancy Prevention Program changes.
- HHS cited seven public health studies in grant solicitations, with two studies fully fabricated and three misattributed, marked as AI-generated citations by the court.
- The injunction blocks HHS's move to restrict funding to abstinence and 'body literacy,' excluding comprehensive sex education.
- A coalition including Minnesota and Washington counties, a sex ed nonprofit, and Planned Parenthood filed the lawsuit in July 2026.
On September 1, 2026, U.S. District Judge Christopher Cooper issued a preliminary injunction against the Department of Health and Human Services (HHS), halting its proposed alterations to the Teen Pregnancy Prevention (TPP) Program. Judge Cooper identified that HHS’s grant solicitations cited seven public health studies, of which two were fully fabricated and three misattributed to journals where they did not exist, an issue the court described as "a hallmark of AI-generated citations."
The policy changes sought to restrict TPP funding exclusively to abstinence-based education and "body literacy"—a term referring to the understanding of human anatomy and physiology relevant to sexual health—effectively excluding comprehensive sex education programs. This shift prompted a lawsuit filed in July 2026 by a coalition consisting of counties in Minnesota and Washington, a sex education advocacy nonprofit, and Planned Parenthood of the Heartland. The complaint challenges HHS’s evidentiary basis for limiting the program's funding.
Judge Cooper criticized the agency’s justification, stating, “While abstinence-only education may well be appropriate for some age groups or communities, the agency has offered no explanation or evidence to support its across-the-board mandate that only abstinence strategies be pursued.”
This decision highlights the legal scrutiny governmental bodies may face when employing AI-generated content in official documents without sufficient verification. It raises broader concerns about reliance on AI tools that may fabricate or misinterpret sources in contexts where accuracy directly impacts public policy and funding allocations.
Legal professionals should monitor the evolving standards of accountability and governance as agencies integrate AI technologies, especially when such technologies influence regulatory or funding frameworks.
By the numbers:
- 7 studies cited by HHS in grant solicitations — 2 fabricated, 3 misattributed
- Sept. 1, 2026 — Date of Judge Cooper’s preliminary injunction
- July 2026 — Month lawsuit was filed by coalition challenging HHS
Yes, but: While the ruling addresses the misuse of AI-generated citations, it remains preliminary; the underlying case will further clarify agency responsibilities regarding AI use.
What's next: Oral arguments in the ongoing lawsuit are scheduled for November 2026, which may provide further judicial guidance on AI deployment in federal policymaking.