AI Hallucination Cases

Rather amazingly, plaintiff asked and was granted leave to file a "Case Law Verification Index" that contained further hallucinations.

The plaintiff, who is a lawyer , filed a claim for damages against Ryanair and included citations to several judgments to support his arguments. It was discovered that the plaintiff had used artificial intelligence to search for these judgments and/or draft the claim , and the cited cases "do not exist". The judge strongly condemned this conduct, stating it was improper and that the plaintiff's excuse for filing in haste was not acceptable. The non-existent citations were disregarded, and the court explicitly denied the plaintiff an award of costs (despite partially winning the claim) as a direct result of this conduct.

See report here.

Relator admitted her petition contained citation errors and hallucinated cases after relying on inadequate assistance; court denied leave to amend and dismissed the writs.

Order to show cause is here.

Debtor cited a case the court could not locate; the court noted the citation appears incorrect and may be an AI "hallucination," and referenced other decisions sanctioning reliance on fabricated AI-generated cases.

In an earlier Report and Reccomendations, the court found that significant portions of the plaintiff's filings copied from an AI program included citations to cases that could not be identified in Westlaw and an apparent AI-generated medical report; the court struck the filings and instructed compliance with Rule 11 and practice standards.

The court observed that Mother's briefs included many cited authorities that do not exist and that one statute she cited was not on point. Mother admitted using Chat GPT to prepare pleadings. The court affirmed the trial court's orders and warned litigants to verify AI-generated citations.

The appellant admitted using generative AI to prepare her Amended Grounds of Appeal. The Appeal Panel and respondent identified numerous incorrect or inapplicable citations and legal propositions generated by AI; the panel noted the appellant failed to verify those citations contrary to NCAT Procedural Direction 7 and treated the Amended Grounds as unreliable in parts when assessing whether questions of law were raised.

Show Cause Order here. Law firm explained what happened here.

In the ultimate order, the court noted that "In terms of competence, the threat to attorneys using generative artificial intelligence platforms powered by large language models is two-fold. First, danger exists that the attorney does not understand how the technology functions, believing that the output is real instead of “realistic-looking."."

In finding that the law firm acted with integrity, the Court noted with approval that it had repaid the other side's fees, to the tune of 55,721.2 USD.

Lawyer later explained that he had "used Microsoft’s Copilot for its editing functions in an effort to review and improve the draft document by fixing grammar, spelling, and improving badly phrased sentences" - not for legal research.

On November 12, 2025, the court resolved the Show Cause proceedings without formal sanctions.

" Moreover, we take this opportunity to caution practitioners of this Commonwealth on the submission of briefs or citations without confirming their accuracy and the correctness of the resulting analysis. The abject failure to conduct due diligence when making arguments to the Court greatly impacts the profession and undermines confidence in the skills and knowledge necessary to practice as an attorney. Failure to verify substantive legal citations prior to submission to this Court is not only in derogation of the RAP, but also violates the attorney's ethical responsibilities. See Supreme Court Rule 3.130(1.1).

Mistakes occur. Oversights happen. Those types of inadvertent errors we could absolve. However, purposelessly submitting a brief to a Court of law without confirming that the cited case law even exists is an affront to the dignity of the Court system, the legal profession as a whole, the judiciary, the client, and the public at large."

The judge tested Microsoft M365 Copilot and found it produced fabricated Court of Appeal citations for Qdime and an incorrect citation for Gala Unity. The Tribunal concluded Mr Gurvits used AI-generated research without adequate verification; the AI outputs were rejected as unreliable. The Tribunal criticised the conduct but did not impose professional disciplinary sanctions; remedial orders in the case related to service charge and fee reimbursement.

The Board found over 70% of citations in Huffman's reply brief inaccurate, including fabricated cases, misattributed reporter citations, cases that did not support cited propositions, and incorrect or non-existent transcript/Rule 4 citations. Counsel admitted using AI to generate portions of the brief. The Board treated the motion as one for Rule 11-type sanctions and struck the reply brief; it emphasized attorneys' duty to verify AI-generated content.

The Court found 28 false or misleading citations across 11 pleadings (14 fabricated, 14 erroneous/misquoted). Mr. Howie admitted use of ChatGPT and failure to verify citations. The Court applied Rule 11(b) and its AI framework (verification, candor/correction, accountability) and imposed sanctions and restitution. Fines of 3,000, 2,000, and 1,000 USD on individual attorneys, plus opposing party's costs and fees,