AI Hallucination Cases

The Father, acting pro se, relied in written submissions on numerous authorities (particularly about ASD) that HHJ Bailey identified as not genuine and apparently AI-generated; the court treated this as part of poor litigation conduct and awarded costs against him.

After being ordered to show cause, Counsel submitted a brief that compounded the issue.

"The Court was dismayed to see that some of the discussion/analysis provided by Mr. Chinweze appeared to be from artificial intelligence generated case summaries which was indicated by the inclusion several times of the statement that "Some case metadata and case summaries were written with the help of AI." Another indication that the paragraphs of discussion/analysis were copied from another source or were generated by AI is that several paragraphs have superscript for footnotes, but no footnotes are found on any page of the Appendix."

The court then cited this database to highlight the growing problem of hallucinations in legal cases.

Counsel was ordered to pay 1,000 USD to the Lawyers Fund for Client Protection, and he was referred to the bar authorities.

The debtor, Ryan Lashon Ford, appeared pro se in a bankruptcy case and was found in civil contempt for failing to comply with court orders. The debtor cited non-existent legal authorities, such as 15 U.S.C. § 114, which the court identified as fabricated citations.

The court also encouraged, but did not require, Plaintiffs' counsel to complete a CLE course on LLMs and legal ethics.

In a response, counsel identified the tools used to write the faulty brief, suggested he would withdraw from the case, and offered to "self-report this incident and offer corrective measures" to the bar authorities.

Rather ironically, plaintiff was seeking to disqualify law firm Butler Snow (of Johnson v. Dunn fame) from the case.

"Deleman admits to using generative artificial intelligence (“AI”) tools while draftinghis filings and has filed certificates of use with the Court. (Doc. 29-2; Doc. 30-2; Doc. 31-3).During the Court’s July 23, 2025, hearing, Deleman appeared to be relying entirely on AIgenerated arguments and citations. The Court directly asked Deleman whether he read thecases he cited and he claimed he did. Given that Deleman did not appear to encounter anexplanation of the term “consideration” while preparing for the hearing, the Court is skepticalof this claim. The Court reminds Deleman that false representations to the Court may warrantsanction."

"Chandra’s self-represented status does not permit him to submit information blindly. Likeevery other person who appears before the court, he has an obligation to confirm that arguments andcase law submitted to the court are supported by existing law, and a failure to do so is sanctionable."

The court imposed a fine of 1 salário-mínimo for bad faith litigation and notified the OAB for potential professional sanctions.

Timothy J. Beall, a pro se litigant, admitted to using AI software to generate filings with fabricated citations. The court did not impose monetary sanctions but required Beall to include a written representation of citation accuracy in future filings. The court acknowledged the challenges faced by pro se litigants but emphasized the responsibility to ensure accurate legal citations.

In their Response, Counsel confessed to the use of AI tools in their Response to the OSC.

(As recounted by Above the Law, the law firm involved quickly deleted a recent post they made about using AI.)

In the Order, the judge prefaced her findings by noting that "Even in cases like this one, where lawyers who cite AI hallucinations accept responsibility and apologize profusely, much damage is done. The opposing party expends resources identifying and exposing the fabrication; the court spends time reviewing materials, holding hearings, deliberating about sanctions, and explaining its ruling; the substance of the case is delayed; and public confidence about the trustworthiness of legal proceedings may be diminished."

The court further reasoned that "At the threshold, the court rejects the invitation to consider that actual authorities stand for the proposition that the bogus authorities were offered to support. That is a stroke of pure luck for these lawyers, and one that did not remediate the waste and harm their misconduct wrought. Further, any sanctions discount on this basis would amplify the siren call of unverified AI for lawyers who are already confident in their legal conclusion. This court will have no part of that."

It added that: "Likewise, the court rejects the invitation to consider that the involved lawyers and firm have been deeply embarrassed in media reports. For many very good reasons, courts traditionally have not relied on the media to do the difficult work of professional discipline, and this court is not about to start."

In a letter, the defendant pointed out the errors in the Opinion - prompting the judge to withdraw it through a minute order, without offering a rationale.

The judge later issued a new opinion, without mentioning or addressing what had happened.

Counsel admitted his law clerk used AI tools and failed to verify the citations. The court imposed a $2,500 monetary sanction, required completion of CLE in ethics/legal technology, and ordered service of the sanction order to the plaintiff.

Case involved counsel sanctioned in Coomer v. Lindell. The court here noted:

"The fact that Kachouroff and DeMaster corrected some of their misrepresentations before the court or the defendants identified them would ordinarily mitigate their conduct. But the reality is that Kachouroff and DeMaster acted only after the Colorado District Court in Coomer v. Lindell noted similar misconduct. That so many misrepresentations persist supports the inference that counsel’s conduct was not mere negligence but an intentional effort to mislead the court."

In her response to the OSC, Counsel disclosed the Westlaw tools she had used.

The pro se litigant filed an appeal with the Supreme Court of the Commonwealth of the Northern Mariana Islands, which was dismissed with prejudice due to his repeated violations of court rules, including citing non-existent legal authorities and hallucinated cases. The court found that Zajradhara's filings were replete with fabricated citations, such as non-existent cases and misrepresented precedents. Despite being warned, plaintiff continued to submit documents with false legal references and engaged in unprofessional conduct, including personal attacks against opposing counsel and court staff. As a result, the court declared him a vexatious litigant, restricting his ability to file future litigation without express permission from the Chief Justice or Presiding Judge. No monetary penalty was imposed due to his financial circumstances.

" Further, the court is aware that relator's brief includes citations to non-existent cases. Relator is put on notice that she may face sanctions in the future for citing false caselaw. See Tex. R. App. P. 52.11(a) (“On motion of any party or on its own initiative, the court may--after notice and a reasonable opportunity to respond--impose just sanctions on a party or attorney who is not acting in good faith as indicated by any of the following... filing a petition that is clearly groundless.”) "

Counsel was sanctioned for citing eight nonexistent cases in briefs filed on behalf of his client, in an appeal concerning the termination of parental rights. The court found that Counsel violated Illinois Supreme Court Rule 375 by submitting fictitious case citations generated by AI without verification. As a result, he was ordered to disgorge $6,925.62 received for his work on the appeal and pay an additional $1,000 in monetary sanctions. The court also directed that a copy of the opinion be sent to the Illinois Attorney Registration and Disciplinary Commission.

Story here, including a link to the original order full of hallucinations.

Judge did not explain why he withdrew the original order and replaced it with a new one, prompting counsel to move for a preservation of the record and an explanation (motion).

The judge dismissed the motion on 1 August 2025, stating: "The original version of that Order contained clerical errors referencing improper parties and factual allegations. The Court promptly corrected those errors and issued an Amended and Corrected TRO on the docket. [...] The Court corrected the record, notified the parties, and the corrected TRO is the controlling order. No further explanation is warranted."

"The first reason I issue sanctions stems from [Counsel]'s claim of ignorance—he asserts he didn't know the use of AI in general and ChatGPT in particular could result in citations to fake cases. Mr. Nield disputes the court's statement in Wadsworth v. Walmart Inc. (D. Wyo. 2025) that it is "well-known in the legal community that AI resources generate fake cases." Indeed, [Counsel] aggressively chides that assertion, positing that "in making that statement, the Wadsworth court cited no study, law school journal article, survey of attorneys, or any source to support this blanket conclusion."

I find [Counsel]'s position troubling. At this point, to be blunt, any lawyer unaware that using generative AI platforms to do legal research is playing with fire is living in a cloud."

[...]

"If anything, [Counsel]’s alleged lack of knowledge of ChatGPT’s shortcomings leads me to do what courts have been doing with increasing frequency: announce loudly and clearly (so that everyone hears and understands) that lawyers blindly relying on generative AI and citing fake cases are violating Bankruptcy Rule 9011 and will be sanctioned"

Counsel submitted a response to an Order to Show Cause that included a false quote attributed to none other than Mata v. Avianca, a case about hallucinations.

According to local press, counsel was called to a special hearing to discuss the citation to an hallucinated authority.

Later on, opposing counsel scoured the full docket and found multiple instances of hallucinations, including for routine, uncontested motions. On July 29, 2025, they filed for sanctions (motion available here; see story here).

Folllowing a show cause order, Counsel admitted to having used the tool together with Google Search, and explained that, although he was aware of the issues with AI models like ChatGPT, he said he did not expect this tool to fall into the same issues.

In May, the court asked Counsel to show cause why they should not be sanctioned for filing briefs with hallucinations - especially since they continued filing hallucinated submissions after being warned about it.

In their Answer, Counsel revealed that "specific citations and quotes in question were inadvertently derived from internal draft text prepared using generative AI research tools designed to expedite legal research and brief drafting".

In the Order, the court noted that Counsel "was not candid to the Court when confronted about his use of AI, stating that some of these documents were “prepared under time constraints,” when he had nearly two more weeks before the deadline to submit his responses." The judge was also unimpressed by Counsel's attempt to shift the blame to a paralegal.

Plaintiff's Counsel cited nine non-existent cases in a response to a motion to dismiss, which were generated using AI software. The court found this to be a violation of Rule 11, as the citations were not checked for accuracy. Counsel admitted the error, apologized, and explained the circumstances, including staff transitions and the use of AI. The court imposed a $1000 sanction, required Mr. Braddy to attend a CLE course on AI ethics, and ordered reimbursement of Defendant's attorney fees and costs.

The submitted document was also an unedited draft containing instructional comments and placeholder text (e.g., "It is recommended to consult with a lawyer...") characteristic of an AI-generated template. The court held:

"The defendants' counsel rightly argued that referring to non-existent citations is particularly burdensome, because it causes a waste of time in checking all the references. Therefore, in light of both the severity of these imaginary citations and the contempt shown by the document as a whole, it is necessary to award exemplary costs in favor of the defendants, as well as costs to the state treasury. " (translation by Gemini 2.5 Pro.)

The Respondent, a self-represented litigant, used generative AI tools for drafting and preliminary research, leading to the citation of a non-existent case, 'Fontaine v Canada, 2004 FC 1777', in his Motion Record. The Court found this to be a fabricated citation, and the (allegedly) intended citation pointed to an irrelevant case.

The court further pointed out that the Respondent had already been caught fabricating citations in a previous proceeding. Despite acknowledging use of AI, the respondent had also failed to provide the declaration on this point required by the AI Practice Direction. The Court ordered the removal of the Motion Record from the file. Costs of $500 CAD were awarded to the Applicant.

Counsel confessed having Claude Sonnet 4 to draft a legal submission, which included fabricated quotations from legal authorities. Counsel said he was pressed by time.

After holding that there is "no reason to distinguish between the submission of fabricated cases and the submission of fabricated quotations from real cases. In both postures, the attorney seeks to persuade the Court using legal authority that does not exist", the court held that Mr. Desmarais had violated Rule 11 of the Federal Rules of Civil Procedure by failing to verify the accuracy of the AI-generated content. Counsel was found to have acted in subjective bad faith, as he was aware of the potential for AI to hallucinate legal citations and failed to take corrective action even after the government pointed out the errors.

The court imposed a $1,000 monetary sanction and required Counsel to complete a CLE course on the ethical use of AI in legal practice and notify his client of the issue.

Counsel tried to blame the software he used, as well as an intern; court was unimpressed

Prior Order to Show Cause available here.

After reviewing - and dismissing - the factual allegations made by Counsel, and noting that they had submitted errata in parallel cases (dealing with other fabricated citations), the court swiftly concluded that they "have violated Rule 11 because they were not reasonable in certifying that the claims, defenses, and other legal contentions contained in Defendants’ Opposition to Motion in Limine [Doc. 283] were warranted by existing law or by a nonfrivolous argument for extending, modifying, or reversing existing law or for establishing new law."

Both counsel were sanctioned with a 3,000 USD fine, payable to the court.