A federal court fined a lawyer $6,000 for three briefs full of AI-fabricated citations

The problem

“A lawyer got fined for using ChatGPT” is now a recurring headline, and the specifics blur together. This case is one of the most-cited early US examples, and the record is precise: an attorney relied on generative AI to draft briefs, filed fictitious case citations three separate times, and was sanctioned under Rule 11.

What was built

Nothing was built. This is an AI-governance honest-negative: the “system” is an unnamed generative-AI tool used as a drafting aid with no verification step. Rafael Ramirez, defense counsel for HoosierVac LLC in an ERISA dispute brought by the Mid Central Operating Engineers Health and Welfare Fund, filed “on three separate occasions, briefs that contained citations to non-existent cases” (source). He admitted he had relied “on generative artificial intelligence tools that produced fictitious case citations,” did not dispute that he failed to verify the generated citations, and did not dispute that his conduct violated Federal Rule of Civil Procedure 11 (source). The three offending briefs were docket entries 39, 52 and 65 (source).

The outcome

The magistrate’s recommendation: $15,000. Magistrate Judge Mark Dinsmore recommended a $15,000 sanction, $5,000 for each of the three briefs, noting the attorney’s “professed ignorance of the propensity of the AI tools he was using to ‘hallucinate’ citations” (source).

The sanction actually imposed: $6,000. On May 28, 2025, District Judge James Patrick Hanlon adopted the report and recommendation in part and concluded that “a sanction of $6,000 ‘suffices to deter repetition of the conduct’” (source). Ramirez was “personally sanctioned $6,000” and ordered to pay the Clerk of the Court by July 10, 2025 (source). The $6,000 figure is independently reported by Bloomberg Law (source).

A separate disciplinary reference. The court’s order records an order of reference of the matter to the Indiana attorney disciplinary commission (source). The underlying conduct, three briefs with fabricated citations, was also reported by Law360 (source).

How this was verified

The origin is a published federal court order, self-authenticating, so no confirmation call is booked against the court. The order text was re-read live on 2026-08-09: the $6,000 sanction, the $15,000 recommendation reduced to it, the three offending filings (Dkt. 39, 52, 65), the July 10, 2025 payment deadline and the “suffices to deter repetition of the conduct” language all appear in it. The $6,000 is independently corroborated by Bloomberg Law; the conduct by Law360; the magistrate’s $15,000 recommendation by TechSpot. The specific AI product used was not named in the order and is asserted by no claim here. Law360’s report is behind a paywall, so it is used only for the conduct, which the primary order already establishes, not as the load-bearing source for any figure.

Related case files

The same Rule 11 mechanism drives Wadsworth v. Walmart, where counsel at a national firm drew a $5,000 sanction for hallucinated citations, a close read on both the rule invoked and the sanction size. In OTG v. Ottogi a court likewise set a Rule 11 figure below the headline exposure for AI-fabricated citations, showing how judges calibrate the deterrent. Gardner v. Combs lands on the same $6,000 amount for a fabricated citation, useful for seeing whether courts are converging on a going rate.

Sources

Cited in this case file. Tier 1 = the federal court order (primary); Tier 2 = independent press naming the parties. Each source was checked against the live page on 2026-08-09.

1. U.S. District Court, S.D. Indiana, “Order on Sanctions, Mid Central Operating Engineers Health and Welfare Fund v. HoosierVac LLC,” No. 2:24-cv-00326-JPH-MJD, Dkt. 121, 2025-05-28 (Tier 1, primary court order). https://websitedc.s3.amazonaws.com/documents/Mid_Cent._Operating_Engrs_Health_v._Hoosiervac_S.D._Indiana_USA_May_28_2025.pdf
2. Bloomberg Law, “Lawyer Sanctioned $6,000 for AI-Generated Fake Legal Citations,” 2025 (Tier 2, independent press on the $6,000). https://news.bloomberglaw.com/employee-benefits/lawyer-sanctioned-6-000-for-ai-generated-fake-legal-citations
3. Law360, “Texas Atty Dinged For AI-Generated Fake Citations In Briefs,” 2025 (Tier 2, independent press on the conduct; paywalled, not load-bearing for any figure). https://www.law360.com/articles/2301530/texas-atty-dinged-for-ai-generated-fake-citations-in-briefs
4. TechSpot, “Lawyer faces $15,000 fine for using fake AI-generated cases in court filing,” 2025 (Tier 2, independent press on the magistrate’s recommendation). https://www.techspot.com/news/106915-lawyer-faces-15000-fine-using-ai-generated-fake.html