Continue below for the international tracker ↓You are on the international page, bringing together judicial comments from around the world. The UK links above open a separate, more detailed tracker.
You are here · The international tracker
International AI hallucination cases tracker
The courts’ own words, from around the world.
Selected judicial quotations on AI hallucination cases and fabricated/false citations from judgments and decisions across jurisdictions.
I bring these comments together to understand how courts are responding, what concerns they express and how their approach to AI in legal proceedings is developing.
This collection forms part of Matthew Lee’s lecturing and teaching in law and writing and editing legal articles and reports. It is communicated solely in connection with promoting or advertising his practice at Doughty Street Chambers. Nothing here is legal advice. No Public or Direct Access. All instructions are through his clerks.
The international reading collection
International AI hallucination cases: search the quotations
Use the search box to find a case name or words in the quoted passages. Browse the table in pages, or clear your search to return to the collection. The three public columns are Date, Case and Judicial Quotes.
This collection brings together judicial comments about fabricated/false citations and related concerns. It includes wider observations about AI in legal proceedings, even where a decision does not establish an AI hallucination.
Read each passage in its original context. A judgment may record an allegation, quote another decision or repeat a party’s submission. Selected extracts may contain indicated omissions or redactions. Independently check the original source and any later developments.
| Date | Case | Judicial Quotes |
|---|---|---|
| 3 Sep 2026 | Douglas v Deutsche Bank National Trust Company | "Our intent is not to discourage attorneys from using AI—indeed, we have reached a point where attorneys eschew at least an understanding, if not use, of AI at their own peril—but to emphasize the importance of doing so discerningly and responsibly. To put a finer point on it, “the use of artificial intelligence must be accompanied by the application of actual intelligence in its execution.” |
| 2 Sep 2026 | Gadiya v Union of India | "At this juncture, we may also add a comment on the use of Artificial Intelligence in adjudicatory processes. While this issue incidentally falls upon us to be considered, use of AI in the process of adjudication is not foreign. Quite the contrary, this Court has recently released draft Regulations for Use of Artificial Intelligence in Courts, 2026, calling for comments and suggestions. Notwithstanding that the regulations are yet to be finalised, use of AI as an assistive tool to speed up the decisionmaking process cannot be denied. However, a note of caution must be sounded: assistance can never be substituted for adjudication. AI may well serve as training wheels but entrusting it with the pilot’s seat would be both imprudent and dangerous." |
| 25 Aug 2026 | SRA v Kumar | "Mistakes happen. Not every mistake will be a matter serious enough to engage the Regulator or this Tribunal. In his response to the Rule 14 Statement, the Respondenthad categorised his wrongdoing as “technological negligence”. His insistence, in his Rebuttal Statement of 28 July and in oral submissions, that he ‘mistakenly believed’that AI was a “lawful drafting aid” reflected a persistent lack of insight into his misconduct. He blamed having been out of legal practice and being, “unaware of recent regulatory warnings or procedural directives regarding AI research tools”. The Tribunal found that the Respondent had embarked on the use of AI without giving any proper thought to the consequences of so doing. He had failed to take proper steps to ensure his work output was accurate or to satisfy himself that the electronic tools hewas seeking to use were reliable. Having done so, the Respondent’s culpability was very high because it was repeated after errors had been pointed out and explained. Rather than take active steps to rectify his mistakes, the Respondent repeated them. The Tribunal gave very substantial weight to the repeated nature of the Respondent’s misconduct." |
| 31 Jul 2026 | Higgins v State of Texas | "During the ensuing colloquy between the trial judge and defense counsel, the trial judge stated that defense counsel’s memorandum misstated the holdings of several cases, whereupon defense counsel admitted that the memorandum was produced by artificial intelligence and he did not read the cases cited therein. The trial judge overruled defense counsel’s objection and admitted the items into evidence." |
| 31 Jul 2026 | The JAAW Group LLC | "As decisions of our Office have explained, while there is nothing inherently wrong with the proper and competent use of AI tools in the legal arena, this evolving technology has many glitches--including hallucinations--and must only be used with close, careful supervision, fact-checking, and citation-checking. KE System Servs., Inc., B-423881 et al., Dec. 22, 2025, at 6. In order to satisfy our statutory mandate to resolve protests and to maintain our role as a meaningful, efficient protest forum, we expect all parties to prepare and present their cases carefully and diligently. Wolverton Prop. Mgmt., LLC-Recon., B-415295.4, June 6, 2018, at 3. This expectation applies to both protesters represented by counsel and pro se protesters proceeding without counsel. See KE System Servs., Inc., supra. The use of AI tools to draft or assist in drafting legal filings can result in the citation of non-existent decisions, such that reliance on those tools without review for accuracy wastes the time of all parties and of GAO. Raven Investigations & Sec. Consulting, LLC, B-423447, May 7, 2025, at 4. As such, our Office necessarily reserves an inherent right to dismiss any protest and to impose sanctions against a protester, where a protester’s actions undermine the integrity and effectiveness of our process. Id." |
| 31 Jul 2026 | Tov Realty LLC v Suarez | "In arriving at this decision today, we do not reject the use of new and improved technologies in the legal profession. Instead, we emphasize that all legal professionals must be acutely aware of the risks and dangers that accompany these technologies, and attorneys must take all steps necessary to ensure that such innovations are utilized with the utmost care and attention to their professional obligations. Indeed, the judges of the Superior Court have recently adopted a new rule of practice addressing the risk that “generative AI can create inaccurate factual and legal information,” and requiring that “any person who uses generative AI in the creation or editing of any document filed with the court . . . independently verify all citations, legal authorities or evidence produced by generative AI.” |
| 31 Jul 2026 | FJL v TGO | [41] I could find no instances in the United Kingdom, Canada, Australia or the United States where a Judge has used AI to reason or determine the outcome of a judicial decision. [42] The most widely cited example is Lord Justice Birss of the Court of Appeal, who publicly disclosed that he had used ChatGPT to produce a summary of an area of law while drafting a judgment. He described the tool as “jolly useful” and indicated that it assisted with summarisation rather than decision-making.10 Importantly, Lord Justice Birss expressly distinguished between using AI as a drafting aid and allowing it to decide legal issues. The AI-generated material was treated as a research or summarisation tool and remained subject to judicial verification. |
| 31 Jul 2026 | Santana v Hardy et al | "Whether this Court has adopted an individual rule regarding the use of AI or not, Rule 11 requires parties to certify that their legal contentions “are warranted by existing law.” Fed. R. Civ. P. 11. Even though he is proceeding pro se, Plaintiff “is required to comply with the Local Rules and Federal Rules of Civil Procedure.” McClellon v. Rickard, 2026 WL 686499, at *9 (S.D.N.Y. Mar. 11, 2026). Accordingly, whether the deficiencies in Plaintiff's filings stem from his use of AI or some other source, the Court cautions him that future submissions to any court must contain only accurate citations and representations. If they do not—because of AI hallucinating case citations or otherwise—he may be sanctioned, id., although the Court declines to do so now." |
| 31 Jul 2026 | Al-Ali v CVS Pharmacy | "The Court warns Plaintiff that, in the future, even one factitious citation will not be tolerated and that she may be sanctioned if future filings contain factitious citations, including but not limited to striking of filings or monetary penalties. The Court warns Plaintiff that factitious citations include case citations to nonexistent cases, but also case citations that, while real, do not stand for their asserted proposition. Plaintiff has an obligation to ensure that she does not submit any factitious citations in the future." |
| 27 Jul 2026 | Lichfield v Kubler, Netflix | "Here, the errors are not as egregious. Although we assume the use of generative artificial intelligence led to the misrepresentations of legal authority, the brief here had significantly fewer errors than in Amarsingh. But, on the other hand, generative artificial intelligence is no longer a novel tool. Attorneys who choose to use it must be willing to also apply “actual intelligencein its execution.” Id. at *6 (quotation omitted). We are further concerned that the Reply Brief did not appropriately remedy the misstatements of law. After briefly acknowledging “several citation errors” and contending that they “have been corrected,” counsel went on to attack the opposing party. Reply Br. at 6. He contended that “[t]he more significant issue” was “mischaracterizations of controlling authority” in the Response Brief. Id. But those purported mischaracterizations were grounded in the words of cases that indeed exist. It is the court that decides which issues are “more significant.” And the errors in the Opening Brief are more significant in this case" |
| 24 Jul 2026 | Lodhia v Twelve Trees et al [2026] EWHC 1889 (KB) | "...did not explain the 22 April documents which he sent to rradar, I asked him about them at the hearing. He said that all of the false evidence and documents were created by AI and were AI hallucinations. He provided a more detailed explanation in his 27th witness statement which was essentially to the same effect i.e. that he had mistakenly sent the wrong documents, AI having generated the false ones without his knowledge and without him realising what had happened. His 27th witness statement includes the detailed prompts which he says he put into the AI tool and exhibits from his output folder, but his explanation requires the court to accept that, having been told on 20 April that his oral criminal law assessment was on 12 May, on the following day he genuinely believed that it was on 6 May. It also requires acceptance that the AI tool generated documents which were false in other respects without any instruction or request from him to do so, that he then mistakenly sent off the wrong versions twice and that he did so without reading them and realising that they were false at any point before Steyn J's Order..." |
| 24 Jul 2026 | Ledoux v Outliers Inc | “[A]ny sanction imposed must be proportionate to the offense and commensurate with principles of restraint and dignity inherent in judicial power.” Oneto, 808 F. Supp. 3d at 980 (quoting Zambrano v. City of Tustin, 885 F.2d 1473, 1480 (9th Cir. 1989)). Courts handling similar cases tend toward leniency where attorneys are forthcoming about their misuse of artificial intelligence and take significant steps to correct their mistakes. See, e.g., Dodge v FirstService Residential Ariz. LLC, No. CV-24-01550-PHX-SMM, 2025 WL 3653164, at *3 (D. Ariz. Dec. 17, 2025) (imposing no sanctions where an attorney “openly and honestly admitted to his improper use of AI, including the submission of hallucinated citations” and “took the initiative to self-report to the State Bar for his improper use of AI and contend[ed] that he [was] reviewing all other pending matters”)." |
| 24 Jul 2026 | Cole v Hobby Town Unlimited | Furthering the Court's concern is the fact that hallucinated cases are a telltale sign of AI usage. AI usage is of special concern to the Court because “[i]t promotes cynicism about the legal profession and the American judicial system.” Mata v. Avianca, Inc., 678 F. Supp. 3d 443, 448 (S.D.N.Y. 2023). In addition, “[w]hile the court takes time to investigate, other cases may be disrupted or deprived of judicial attention.” Virgil v. Experian Info. Sols. Inc., No. 1:25-cv-01641-MPB-MJD, 2026 WL 401066, at *4 (S.D. Ind. Feb. 12, 2026). Finally, and perhaps what should have been most troubling to Plaintiff's counsel here, is that, as a result of AI usage, “client[s] may be deprived of arguments based on authentic judicial precedents.” Mata, 678 F. Supp. 3d at 448. |
| 24 Jul 2026 | Graves v Pacif Gas and Electric | “At the hearing on these motions, the undersigned admonished Plaintiff about the danger of using Artificial Intelligence tools for legal research and writing and Plaintiff's obligation to ensure that all legal citations are legitimate. Future citations to non-existent cases may subject Plaintiff to an order to show cause as to why she should not be sanctioned…” |
| 23 Jul 2026 | Graves v Graves | "As a final matter, we admonish … for citing non-existent case law in his brief but decline ….’s request to dismiss the appeal as a sanction." |
| 23 Jul 2026 | Hagan v Black Diamond Experts Et al | "The Court takes seriously allegations of the improper use of artificial intelligence. Nonetheless, the Court concludes that, given the unusual circumstances described in the Stipulated Motion to Stay regarding … counsel’s asserted “incapacitation,”40 further pursuit of this issue at this time would not be fruitful. Generally, however, the Court admonishes counsel to be vigilant with respect to artificial intelligence. If any improper use of artificial intelligence is determined to have occurred in the future, sanctions may be imposed." |
| 23 Jul 2026 | Porfirio v Mullin | “But the Court notes that the brief’s citation to nonexistent quotations and cases as well as its misstatements of law bear the hallmarks of reliance on an artificial intelligence tool. See Williams v. Cap. One Bank, N.A., No. 15 24-cv-2032, 2025 WL 843285, at *7 (D.D.C. Mar 18, 2025) (“Courts have recently seen increasing reliance on artificial intelligence in legal proceedings, leading to the use of nonexistent citations in court documents.”). And if this case somehow returns to this Court, the Court may inquire further into the preparation of the opposition brief.” |
| 28 May 2026 | Blanes v. Stellers Jay Academy Inc. | I find these authorities are likely “hallucinations,” meaning false or misleading results generated by artificial intelligence. |
| 27 May 2026 | Innocent v. Meraki Installers, LLC | "This appeal follows a final judgment entered in favor of the Appellee in a lawsuit for breach of contract and foreclosure of a construction lien. We affirm without comment on all issues. We "write briefly, however, to address concerns regarding multiple quotations from the trial transcript contained in Appellant’s initial brief that do not appear to be foundanywhere in the actual trial transcript or otherwise in the Record. In consideration of the arguments made on this point in the parties’ briefs, as well as during oral argument, we refer this matter to the Florida Bar for appropriate action. |
| 27 May 2026 | Eclectic Synergy, LLC v. Seredin | "We denied the “emergency” motion for stay,1 and issued an order requiring counsel to show cause why sanctions should not be imposed because the petition appears AI-generated and raises frivolous arguments, misstates the law, and cites non-existent caselaw." |
| 27 May 2026 | Robert Hinton Avery v. Danielle Beauzil | "This argument misstates the law, and the cited cases do not support the argument...The law does not support petitioner’s argument at all. These types of AI-hallucinated arguments are just as egregious as a citation to non-existent law...." |
| 27 May 2026 | Williamson v. Meyers (derivatively on behalf of Equinix, Inc.) | "In reviewing Plaintiffs briefing, the Court identified at least one existence where Plaintiff had cited to a non-existent authority and brought this issue to Plaintiffs attention. Afterinvestigating, Plaintiff conceded that this fictitious citation was most likely the result of artificial intelligence ("Al"). The Court has not considered any of Plaintiffs contentions, to the extent that they rely on fabricated authority, and will consider whether any other action is necessary." |
| 27 May 2026 | Christ Apostolic Temple, Inc., Dwight Reed and Jordan Reed v. Demetrius Sinegal, et al. | "...briefs contain citations to cases that do not exist or do not stand for the proposition for which they are cited. This suggests the possibility that he used generative artificial intelligence to prepare his briefs. If that is the case, we again “stress that self-represented litigants and attorneys alike have a duty to independently verify the authenticity and veracity of all sources and assertions when relying on artificial intelligence tools to prepare trial or appellate court filings." |
| 27 May 2026 | Chen v. The Owners, Strata Plan LMS3094 | "I find these are likely “hallucinations”, meaning false or misleading results generated by artificial intelligence. ...While I havereviewed all of ... submissions and evidence, I have only addressed what is relevant in my decision." |
| 27 May 2026 | State of Oklahoma ex rel. Oklahoma Bar Association v. Reeves | "There is nothing inherently problematic with the use of generative artificial intelligence in preparing legal materials, so long as attorneys abide by their duty to protect client confidentiality and recall their sacred duty to verify the pleadings they sign. The careless use of generative artificial intelligence to "save time" by not independently confirming citations is instead a waste of judicial resources and opposing party's time and money. It also damages the integrity and credibility of the legal system. Human diligence and review is required to ensure content and accuracy of filed documents. Signing pleadings that contain citations from generative AI that have not been verified shows a reckless disregard for the truth and an indifference to accuracy." |
| 18 May 2026 | Anthony Malcolm Cork & Anor v Mark Smith Neutral Citation Number[2026] EWHC 1199 (Ch) | 95. I do not intend to set out the words of the President in Ayinde again but it will be clear from what she said that legal professionals bear ultimate responsibility for their work and cannot outsource the process of legal research or of legal reasoning to an AI. It is a tool to be used with caution. AI has the potential to be wholly unreliable. AI may of course provide a jumping off point for research and legal reasoning but it does not, at least at present, do away with the need for proper research and thought on the part of a legal professional, even a very junior legal professiona |
| 16 May 2026 | Jeffs v London Borough of Lewisham | “32…(b) The Applicant has misunderstood the law and relied on inaccurate statements which the Tribunal understands to have originated from AI hallucinations which included false quotes of legislation and cases that were not relevant or did not exist.” |
| 14 May 2026 | Rodney -v- Gee’z Micro Bar | "It is important that there is a public record of this incident should anything remotely similar occur again involving [redacted firm] or these solicitors. I therefore direct that a transcript of this judgment be prepared at public expense and be published on the Judiciary website. A copy shall be sent to the Solicitors Regulation Authority alongside the witness statements filed by both solicitors and my order of 5 March 2026." |
| 6 May 2026 | Rafique v The Commissioners for HMRC [2026] UKFTT 673 | 100. While litigants in person in the FTT are (entirely appropriately) given a greater degree of latitude in their conduct of the proceedings than represented parties, no person has immunity from sanction when committing a serious breach of court or tribunal rules. That is well demonstrated in Olsen v Finansiel Stabilitet A/S [2025] EWHC 42 (KB) where the High Court only “narrowly and somewhat reluctantly” decided against causing a summons for contempt of court to be issued to two elderly litigants in person who had provided a fabricated citation to the court. |
| 5 May 2026 | Tuyunuklu v Akmal [2026] UKUT 174 | 16. In considering the parties’ arguments I start from two propositions of law. The first is that Rakusen v Jepsen is not authority for the idea that [redacted principle]. The Supreme Court did not say that. From references to AI in the transcript I wonder if the appellant was misled by an AI hallucination. Whatever the reason, his point about Rakusen v Jepsen is without foundation |
| 11 Mar 2026 | Setu Kamal v Tax Policy Associates Ltd & Anor | 184. At a more granular level, the Defendants object to the variety of different addresses, in different jurisdictions, [claimant] has used on his claim form, particulars of claim, application notices and witness statement. They object to his imposition of arbitrary and oppressive deadlines for responses after long periods of delay. They object to a range of features of his correspondence including misapplying the ‘without prejudice’ designation to letters containing no settlement proposals, and the deployment of AI-generated ‘hallucinatory’ references to non-existent cases causing unnecessary effort in cross-checking (there being already a growing jurisprudence of court deprecation of this phenomenon, including in exercise of the Hamid jurisdiction – see R (Ayinde) v Haringey LBC [2025] EWHC 1383 (Admin)). |
| 5 Mar 2026 | Re A, B, C, D (Extension of assessment; Use of AI: hallucinations) | It is the duty to the court owed by [Representative} as a litigant in person to ensure that the cases cited in legal argument are genuine and provide authority for the proposition advance with which I am concerned. She accepts that she did not do so, and apologises. She says that the mistakes were unintentional, and seeks to justify herself by saying the main authority upon which she relied was sent to me (as it was). Having heard from her, and read her submissions, I absolve her of any intention to mislead the court, but remain concerned that [Representative] minimises the seriousness of misleading the court and goes so far as to assert that criticising use of AI risks setting a harmful precedent for disabled litigants in person and will discourage access to justice. |
The reason for this collection
AI hallucination cases: what are courts saying?
My interest in AI hallucination cases is in the language and reasoning of the judgments: how courts describe unreliable material, distinguish suspicion from an established finding, respond to explanations and discuss the proper use of AI.
This public page is deliberately focused on selected passages. Reading them together can reveal shared concerns, different approaches and questions worth exploring as the international discussion develops. The collection is selective and does not measure how often these issues occur or establish a worldwide consensus.
Case names identify the source and allow the passage to be placed in context. My purpose is to study the courts’ comments and contribute to an understanding of AI and legal practice, rather than to name or shame individuals.
Questions running through the judgments
Reliability and verification
What courts say about checking citations, quotations, evidence and legal propositions.
Context and response
How judgments record explanations, uncertainty, responsibility and the court’s response.
The developing role of AI
Comments about legitimate assistance, professional judgment and the limits of automation.
Keeping the words in context
Fabricated/false citations: reading the comments in context
Does an entry prove an AI hallucination?
The distinction matters. A judgment may report an accusation, reproduce a submission, express a concern or make a finding. Those are different things. An extract must not be treated as establishing more than the decision actually says.
What does “fabricated/false citations” cover here?
It is a research label for the subject discussed in the source, which can include fake legal citations, invented authorities, inaccurate quotations, misleading references or concerns about a legal proposition. It does not establish how the material was produced. The wording of each judgment, including any uncertainty about AI, remains central.
Why include comments where AI use is not established?
The purpose is to understand the international discussion of AI hallucination cases and related citation problems. General warnings, disputed explanations and comments about appropriate AI assistance can contribute to that discussion. A relevant judicial comment can therefore justify inclusion without a finding that anyone used AI.
Why retain case references?
They identify the decision so readers can understand its context. A case name is a source reference, not a label attached to a person’s conduct. The purpose of the collection is comparative reading and research.
Why are some passages shortened or redacted?
The public collection concentrates on the relevant comment. Ellipses indicate omissions and square brackets may indicate editorial substitutions or redactions. These should not be read as additional words spoken by the court. The full judgment remains the source for its complete wording, reasoning and outcome.
How does this differ from the UK tracker?
The UK AI Hallucination Cases Tracker offers more detailed UK records, outcomes, sources and research charts. This international page concentrates on selected judicial comments across jurisdictions, including fabricated/false citations and wider observations about AI. A UK decision may also appear here where its passage contributes to that wider discussion.
Does the collection predict what courts will do?
No. It brings together comments for study. Decisions arise in different legal systems and factual settings. Similar language does not necessarily mean the same rule or outcome applies elsewhere.
The research behind the collection
A public collection, supported by more detailed research
I maintain a separate, more detailed personal research tracker containing additional case information, references and research notes. The extracts published here are one part of that work and do not reproduce the full private tracker.
For enquiries about the research, possible access, a proposed addition or a correction, please contact my clerks at Doughty Street Chambers. For a correction, identifying the case, passage and source helps me locate the point.
AI law · Advice · Advocacy
Matthew Lee’s AI law practice
I accept instructions on many AI law matters, including AI issues arising within existing proceedings or another area of practice. My work includes advice, drafting, advocacy and representation.
If an AI legal issue arises, please contact my clerks at Doughty Street Chambers to discuss instructing me.
Contact my clerksContinue exploring
More AI law resources
UK AI hallucination cases
More detailed UK records, outcomes, judicial quotations, source links and figures from the wider research.
Open the UK tracker →Will AI replace lawyers?
Public statements about lawyers, legal work and the future of the profession.
Explore the tracker →International deepfake cases
Proceedings and decisions involving alleged or established synthetic media.
Explore the database →Judicial AI Use Tracker · AI privilege and confidentiality · All AI law trackers and resources
Legal writing by area of practice
The AI Law Reports
Explore AI-related legal developments through familiar areas of legal practice in The AI Law Reports, founded by Matthew Lee alongside Natural & Artificial Law.
Important information
Matthew Lee accepts no responsibility or liability for errors, omissions or loss arising from use of or reliance on these resources. Nothing here is legal advice.
AI tools assist with research, organisation and updates. Extracts, references and other information may contain errors, omissions or inaccuracies and may be incomplete or out of date. Readers must independently check the original decision, its full context and current status and must not rely on this collection when deciding how to act or as a substitute for legal advice.
Inclusion does not establish AI use, wrongdoing or the truth of an allegation. Making an enquiry does not itself mean that instructions have been accepted.
Key Takeaway
This public-facing international tracker records public references to suspected or confirmed AI hallucination cases and other suspected or confirmed incidents of fabricated/false citations in court judgments, tribunal decisions and legal proceedings worldwide. Inclusion in this tracker does not mean that any named person, lawyer, party or court necessarily used AI, was suspected of using AI, acted improperly, or was found to have done so. It means only that the public source records, raises or discusses an issue concerning AI hallucinations, fabricated/false citations, or something similar.
The fuller private tracker, available only by request, contains deeper research including the false citations themselves, whether judges repeated them in judgments, the types of fabricated or false material, whether they concern case law, legislation or something else, who discovered them, whether the issue involved litigants in person/pro se or lawyers, the reasons given, and any aggravating or mitigating features. This tracker is always evolving and cases may be added or removed at any point.
