AI Fake Cite Sanction Reduced To $6,000 From $15,000 By Judge

·

(May 30, 2025, 2:22 PM EDT) -- INDIANAPOLIS — An Indiana federal judge partially adopted a magistrate judge’s recommendation that the court impose sanctions against an attorney for submitting briefing with artificial intelligence-created fake citations, reducing the sanction to $6,000 from the recommended $15,000 while rejecting the attorney’s claim that the reputational damage he incurred as a result of using the fake cite mooted any need for further proceedings.

(Mid Central Operating Engineers Health and Welfare Fund v. HoosierVac LLC, No. 24-326, S.D. Ind.)

(Order available.  Document #46-250604-080R.)

U.S. Judge James Patrick Hanlon of the Southern District of Indiana entered the order on May 28.

HoosierVac LLC uses union vacuum truck operators in its business.  The operators belong to the International Union of Operating Engineers.  The union in turn maintains the Mid Central Operating Engineers Health and Welfare Fund.  The fund and trustee Stephen Scott sued HoosierVac in the U.S. District Court for the Southern District of Indiana, claiming that it violated the union’s collective bargaining agreement by not making timely and full contributions to the employee benefit plan.

Also in the dispute, HoosierVac filed complaints with the National Labor Relations Board region office in Indianapolis against two local chapters of the union.  HoosierVac then moved in the District Court to transfer the NLRB region office actions to the federal court.  The District Court denied the motion, finding that no mechanism existed to make such a transfer.  HoosierVac moved for reconsideration of the ruling in October.  The motion remains pending.

Withdrawn Cite

U.S. Magistrate Judge Mark J. Dinsmore of the Southern District of Indiana notified HoosierVac attorney Rafeal Ramirez that one of the cites in the motion to reconsider appeared incorrect.  Ramirez conceded the error and withdrew the cite.  Magistrate Judge Dinsmore ordered Ramirez to appear in person and show why a sanction should not issue.  Magistrate Judge Dinsmore noted that a more exhaustive search of the documents in the action found two other instances where a cited case could not be found.

At the resulting hearing, Ramirez conceded that he relied on AI to draft the briefs and that he was unaware that AI could generate fictitious citations.  Ramirez said at the hearing that because the AI-produced results appeared legitimate, he did no further investigation.  Ramirez said he had undergone legal education on the topic of AI and conceded that while his conduct did not meet the standards of Federal Rule of Civil Procedure 11(c)(3), Fed. R. Civ. P. 11(c)(3), the mistake was not made in bad faith.

In the report and recommendation, Magistrate Judge Dinsmore noted that various legal tools such as LexisNexis’ Shephardization have made fulfilling the obligations of Rule 11 easy.  “There is simply no reason for an attorney to fail to fulfill this obligation.  Such has been the view for decades,” Magistrate Judge Dinsmore said.

[Editor’s note: Mealey Publications is a division of LexisNexis Legal and Professional, a RELX Group company.]

In recommending that Ramirez be sanctioned $15,000, Magistrate Judge Dinsmore noted that Ramirez apologized and withdrew the cite.  But he said filing documents with the court containing improper cites “falls far short of an attorney’s duty.”

“It is one thing to use AI to assist with initial research, and even non-legal AI programs may provide a helpful 30,000-foot view.  It is an entirely different thing, however, to rely on the output of a generative AI program without verifying the current treatment or validity—or, indeed, the very existence—of the case presented.  Confirming a case is good law is a basic, routine matter and something to be expected from a practicing attorney,” Magistrate Judge Dinsmore said.

Magistrate Judge Dinsmore referred the matter to the chief judge for any further discipline.

Sanction

In a March 7 response to the magistrate judge’s report and recommendation, Ramirez reiterated that he took responsibility for the error, has suffered irreversible reputational harm from disclosure of the error and has taken steps to education himself on responsible use of AI.  As a result, the error is not likely to reoccur, Ramirez said. 

Given that the error occurred without ill will or malice and the reputational damages already incurred, the principle of mootness applies and no further proceedings were necessary, Ramirez argued.

In his order, Judge Hanlon said the precedents on which Ramirez relies for his mootness argument don’t apply.

Ramirez relies on cases where the court could no longer affect the rights of the litigants, Judge Hanlon said.  That is not the case in the action at hand, where a Rule 11 sanction is within the court’s jurisdiction and therefore not moot, Judge Hanlon said.

Judge Hanlon said that taking into account the steps Ramirez already took to educate himself about AI, a $6,000 sanction suffices.

Counsel

HoosierVac is represented by Ramirez of Ramirez Law Office PC in Rio Hondo, Texas.

The fund is represented by Daniel Bowman of Bowman & Vlink LLC in Indianapolis.

(Additional documents available:  Ramirez’s response to magistrate judge’s report and recommendation.  Document #46-250604-079B.  Report and recommendation.  Document #46-250305-042Z.  Order identifying fake cite.  Document #46-250305-047R.  Order to show cause. Document #46-250305-048R.)