AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations)

Ad/Marketing Communication

UK‑based legal commentary and comparative analysis of international case law on AI related legal issues. This AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations) forms part of lecturing/teaching law and writing/editing law articles/reports and is communicated solely in connection with promoting or advertising Matthew Lee’s practice. Not legal advice. Not Direct/Public Access. All instructions via clerks at Doughty Street Chambers.

19 January 2026 Update on AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations)

Thank you for your patience while I update the AI Hallucinations Cases Tracker (AI and non-AI fabricated/false citations).
The latest entries and analyses are being added shortly. In the meantime, enter your email below to be notified when the update is complete and to receive regular insights and case summaries.

What you can see

You will see two things below: the AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations) and a set of Charts/Diagrams. The tracker lists the case reference (neutral citation), a link to the public judgment, and a short extract from that judgment.

What the Charts Show

The charts below give a rough picture of the data I’m currently analysing on AI and non-AI fabricated/false citations. They offer early insights into some of the key questions I’m exploring, including:

• Did the judge cite the fabricated or false citation within the judgment?
• What types of AI and non-AI fabricated/false citations are finding their way into court?
• Are these AI and non-AI fabricated/false citations referring to case law, legislation, or something else?
• Who is discovering the AI and non-AI fabricated/false citations?
• Is this mainly occurring with litigants in person (pro se) or with lawyers?
• What reasons are given for the AI and non-AI fabricated/false citations?
• What features are aggravating or mitigating?

What’s in the Vault

These visuals are powered by my private research, which is not included in the publicly available AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations) below. This private research contains deeper analysis, richer tagging, broader categories, detailed timelines and considerable detail on fabricated/false citations.

I do not publish the underlying databases or research publicly. If you wish to discuss access to my private research, please contact my clerks (details above)

Important Note

This tracker is part of my ongoing research. It takes time to build and is updated frequently as I work through new judgments and revisit the historic. For further analysis of these cases, I would suggest reading my regular commentary on AI Hallucinations (AI and non-AI fabricated/false citations) cases globally. The latest is available [here]. Most cases are from the USA, but this is an international issue.

The aim of this AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations) is to gather broad data about how hallucinations, fabricated citations, whether by AI or otherwise, proven or just alleged, are appearing in courts and reports. As set out above, I am interested in questions such as: Why is this happening? What types of hallucinations are occurring? What reasons are given? What mitigating or aggravating features are raised? And whether judges themselves cite these errors in their authentic judgments? I want to understand the broader international picture to advance AI law, an important area of my legal practice in England and Wales.

Judgments do not all describe these issues in the same way, and they do not always fit neatly within a label. To spot patterns and make study easier, I group what is recorded, advanced, or alleged into broad categories. This means the tracker will not always match the exact wording of the judgment and it may be incomplete or contain errors. For the detail in any particular case, always check the official judgment via the hyperlink. If the hyperlink takes you to one of my legal articles on this blog, you will be able to access the judgment from that legal article.

I also use AI tools and other online resources to help organise and maintain this AI Hallucination Cases Tracker (AI and non-AI fabricated/false citations). For that reason, and because the task involves categorising complex material, you should always verify details against the judgment itself. If you notice anything that looks inaccurate, please let me know and I will review and update.

How To Read the Columns and the Charts

Actor
Based on the judgment, I record whether it was a lawyer, a litigant in person, or another participant connected with the alleged error. Different judgments use different wording, so I seek to place them into consistent groups to make the data comparable.

Hallucination Type
I have grouped the most common types of AI and non-AI fabricated/false citations into eight categories, explained in detail in a separate article. Where something does not fit perfectly, I place it in the closest type so it can be counted and compared across judgments.

Reasons
This records reasons either found by the court or advanced by a party. They are grouped into broad categories such as “time pressure” or “lack of training” so that different explanations can be studied together. Always check the judgment for the precise detail, as reasons are often more complex than the summary category suggests.

Mitigation
This includes both mitigating factors the court accepted and those advanced by a party, such as “acted honestly” or “tried to put it right”. They are grouped so we can see the range of issues raised across cases and compare how courts respond. Specific mitigation is always more nuanced than the summary (always check the judgment for the full detail).

Aggravation
This includes aggravating factors noted by the court or advanced in submissions, such as “repeated incident” or “ignoring guidance”. They are grouped into categories for comparison. Again, for the exact findings, always refer to the judgment.

Cited Fake Case
A simple yes/no marker showing whether the judge cited a hallucinated case/fabricated citation or principle within the judgment itself. Why this issue matters is explained here.

Research Upload in Progress – Thank You for Bearing with Me!

This tracker is being updated in real time, so some parts may look unfinished for a few days, so please frequently check in. There’s also plenty to explore in the meantime:

• The original tracker: still live and available below!
• What are AI Hallucination Cases and the 8 Types: if you’d like to understand what a legal hallucination actually is and the different ways it appears, take a look at the 8 types I’ve identified in this article. These will form the backbone of the upgraded tracker.
• Weekly write-ups (International): every week I summarise the most recent AI hallucination cases. The latest is available [here]. Most case are from the USA, but this is an international issue.
• The UK Position: the UK incidents are reported separately here.
• Popular Posts: read all about this project and popular posts/AI trackers here.
• Check out the FAQ below for common Questions
• Subscribe: I send regular updates from Substack here direct to your email (It’s FREE). Be the first to receive important updates in AI law.

Thank you for your patience – the full launch is coming very soon. In the meantime, dive into the existing resources and see how the field is already evolving.

Key Stats Line & Charts

Total entries: [TBC] Last updated: [22 November 2025].

AI Hallucination Cases Tracker Under Construction

The Older AI Hallucination Cases Tracker below

AI hallucination cases can, and do, reach real courtrooms and have real consequences.

What the AI Hallucination Cases Tracker is for.

Generative-AI tools sometimes invent case citations, statutes, articles or facts. When those errors slip into legal documents the consequences are serious. Sometimes elements get struck out, fines are imposed, lawyers sanctioned and in extreme cases maybe even committal proceedings launched. This page tracks publicly reported decisions and strives to be one reliable index of AI hallucination cases in live legal practice.

How to use the AI Hallucination Cases Tracker

• Each row summarises one decision. You’ll see the country, the date, which AI tool was involved, a short note describing the hallucination, what was done about it or is proposed, and whether the person using AI was a lawyer, a litigant-in-person (self represented) or other.
  
• Click Link in the last column to open a source it is hoped that the links will provide some authoritative source, but if it doesn’t and you are aware of one, please let us know. The best sources are usually a court website, a statutory legal archive such as BAILII, CanLII or CourtListener, or a PDF published directly by a disciplinary tribunal etc.
  
• The table is searchable and sortable: type in the search box to filter by country, tool or sanction, or click any column header to sort.

Why you still need to read the original document
Although I try and check each entry carefully, this page is only a summary. The authoritative version of every decision is the one you must locate yourself. Procedures can move fast: sanctions may be appealed, recommendations can become final, costs can be varied and I may not update in time. Always read the original judgment, order or ruling and you should not rely on this table in practice, publication or teaching.

Keeping the tracker current
I try to add new AI hallucination cases as soon as they are brought to my attention. The page updates automatically because the table is managed in TablePress; whenever a new row is saved, the latest version appears here without extra work. Look for the “Last updated” line beneath the table to know when the most recent changes were made.

Scope
Not all hallucinations are confirmed. In some cases, it was merely alleged. Alleged hallucinations may be included if the allegation itself was significant to the proceeding.

Why the problem matters

1. Accuracy is fundamental. Judges rely on counsel to supply real authorities. AI hallucination cases erode trust in written advocacy and evidence.
2. Professional duties are evolving. Many ethics codes now say lawyers must understand the technology they use. Sanctions in these decisions often reference competence rules. This is discussed elsewhere on this blog.
3. Public confidence is important. When headlines report AI inventing cases, the public questions the legal system’s rigour. Tracking responses across jurisdictions shows how the profession is confronting the risk.

Caution
This tracker is provided for general information only and is not legal advice. The summaries and links do not replace the official judgment, and if you wish to rely on any case you must obtain the authoritative version yourself, direct from the court, tribunal or official reporter as appropriate. Do not rely solely on the information or links shown here. Before taking any action, obtain independent, qualified legal advice.

Last updated: 16 May 2025

Case Name	Country	Date	AI Tool	Hallucination	Judicial Observation (Quote)	Notes	Who Used AI	Link
Kuzniar v General Dental Council (ET 6009997/2024)	UK	20 August 25	ChatGPT/Other	False Citations	“The Claimant conducted the claim unreasonably as described above by referring to the Respondent a large number of nonsensical and in many cases non-existent citations without taking any or sufficient care to check them first. By not doing so she passed the work of checking them to the Respondent to have to do at short notice. My discretion to award costs is engaged. Furthermore, although I did not make any formal enquiry into her financial means, she told me that she has only £2000 in the bank and is struggling to find work as a dentist because of the conditions imposed by the Respondent. However, I decline to award costs because AI is a relatively new tool which the public is still getting used to, the Claimant acted honestly (and furthermore has presented her case honestly to me over the last two days), and she tried to her best to rectify the situation as soon as she became aware of her mistake.” (paras 45-47)	No sanction or costs order due to new tool, honestly, and best efforts to rectify mistake	Litigant in Person	link
Father v Mother [2025] EWHC 2135 (Fam	UK	30 July 25	Unspecified	False Citations	“(16) The F then made a further application on a C2 asking that HHJ Bailey recuse herself on the basis of being biased against him and her not understanding ASD and the impacts of his diagnosis. This came before the Judge on 10 June 2025. In his written application to the court the F referred to a number of previous authorities, in particular relating to ASD. HHJ Bailey realised that many of these cases were not genuine, and the submission appeared to have been generated by Artificial Intelligence (“AI”). In light of the level of recent concern about litigants and lawyers using AI and referring to cases which are not genuine (as reflected in the Divisional Court decision R (Ayinde) v London Borough of Haringey [2025] EWHC 1383), HHJ Bailey referred the case to me as the Family Presiding Judge for the Midlands.” “The F relied upon faked cases without apparently making any effort to check their veracity. It is in my view important to note that the F is someone who is well capable of checking references and ensuring documents are accurate if it is in his interests to do so.”	Judge ordered F to pay the costs	Litigant in Person	link
HMRC v Gunnarsson [2025] UKUT 247	UK	23 July 2025	Unspecified	False Citations	“…In this case, HMRC was put to the trouble of having to investigate the existence of the purported decisions relied upon by the Respondent. Fortunately, they did so. Depending on the circumstances, there may be occasions when the opposing party or the tribunal are not able to discover the errors relied upon. There may be others where an adjournment is required to investigate or address the inaccurate information…”	Warning to all Court users	Litigant in Person	link
Ms (Bangladesh) v SoS for Home Department	UK	1 July 2025	ChatGPT	Unclear if Hallucination	13. We sought clarification regarding this citation and reference and asked for the relevant paragraph of the judgment being relied on. [counsel] was not able to specify this. [counsel] submitted that he understood, having used ChatGBT, that the Court of Appeal in Y (China) [2010] EWCA Civ 116 was presided by Pill LJ, LJ Sullivan LJ and Sir Paul Kennedy. However, the citation [2010] EWCA Civ 116 did not point to the case of Y (China) but to R (on the application of YH) v SSHD. We raised concern about this and referred [counsel] to the recent decision of the President of King’s Bench Division in Ayinde [2025] EWHC 1383 (Admin) on the use of Artificial Intelligence and fictitious cases, and directed him to make separate representations in writing. 14. In his subsequent written representations, [counsel] clarified that Y(China) was a typological error and he sought to rely on R (on the application of YH) v SSHD [2010] EWCA Civ 116 where, when discussing the meaning of ‘anxious scrutiny’ in asylum claims…”	TBC	Lawyer	link
UB v SoS for Home Department	UK	18 June 25	Unspecified	False Citations	The drafter then “inadvertently uploaded the draft version [of the Grounds] rather than the final one”. The Judge accepted the explanation and that solicitors had: “…recognised this seriousness of this issue and has taken commendable steps to ensure it will not be repeated including (i) meeting with the caseworker who drafted the Grounds; (ii) holding a partners’ meeting to discuss adopting an AI policy and assigning the task of finalising an AI policy to a colleague in consultation with an AI professional; (iii) conducting relevant in-house training and issuing interim AI Guidance and (iv) planning for comprehensive staff training by an AI professional….”	Warning	Lawyer	link
(BL O/0559/25)	UK	July 2025	ChatGPT	Type 7-8 potentially	“As identified in Ayinde (including in the Appendix setting out domestic and overseas examples of attempts to rely on fake citations), fabrication of citations can involve making up a case entirely, making up quotes and attributing them to a real case, and also making up a legal proposition and attributing it to a real case even though the case is not relevant to the legal proposition being made (for instance, it deals with a completely different issue or area of law). It is not, however, fabrication to make an honest mistake as to what a court held in a particular case or to be genuinely mistaken as to the effect of a court’s judgment. In any event, it does not matter whether fabrication was arrived at with or without the aid of generative artificial intelligence. I therefore need to consider what if any sanction is appropriate.”	Warning to both	Litigant in Person and Trade Mark Attorney	link
Shahid v Esaam	US	June 25	AI not certain	Fake cases	“We are troubled by the citation of bogus cases in the trial court’s order. As the reviewing court, we make no findings of fact as to how this impropriety occurred, observing only that the order purports to have been prepared by Husband’s attorney, …. We further note that Lynch had cited the two fictitious cases that made it into the trial court’s order in Husband’s response to the petition to reopen, and she cited additional fake cases both in that Response and in the Appellee’s Brief filed in this Court.”	Judge cited fake cases in court order	Lawyer and Judge	link
Jakes v Youngblood	US	June 25	AI use disputed.	False Citations	“These quotations are merely representative of the fabricated statements in [B’s] briefs. There are additional fabricated quotations that the Court does not enumerate in this order. In addition to including non-existent quotations in his briefs, Blackburn also cited cases for propositions that they do not represent. The Court will not recite every time [B] misconstrued a case in his briefs as it believes the above quotations represent the most serious and alarming issues with the documents. Attorneys are permitted to make creative case comparisons and may even stretch existing case law to support their arguments. Nevertheless, advocacy is confined by Rule ll(b) and Pa. RPC 3.3. Attorneys have a duty of candor to the Court. They must make reasonable inquiries under the circumstances to ensure their legal contentions are warranted by existing law. Attorneys may not fabricate non-existent quotations, from case law or the Court’s opinion, and may not cite cases for legal propositions for which they do not stand (or even discuss).”	Further hearing listed 24 July 2025	Lawyer	“Link
Not specified – referred to in Ayinde below.	UK	April 25	Unspecified	Non existent cases.	“That was a case before the County Court … That counsel drew attention to the fact that the application before the judge contained false material: specifically the grounds of appeal and the skeleton argument settled … contained references to a number of cases that do not exist….”	Judge satisfied by assurances given to court by Barrister and Head of Chambers not to take further action.	Lawyer	“Link Link“
Mid Central Operating Engineers HWF v Hoosiervac LLC	US	May 25	Unspecified	Fake cases	“That said, in considering an appropriate sanction the Court takes into account the steps [Lawyer] has taken “to educate himself on the responsible use of AI in legal practice” and adhere to “the highest standards of professional conduct moving forward.” See dkt. 102 at 2. The Court also considers the collateral consequences that [Lawyer] has experienced, and may continue to experience, from having improperly relied on non-existent AI-generated legal citations…so, the court has considered those circumstances alongside its interest in deterring careless or reckless attorney conduct.”	Sanction of $6,000 not accepting recommended $15,000	Lawyer	Link
Versant Funding LLC v Teras Breakbulk Ocean	US	May 25	Unspecified	admitted submission of a wholly fabricated “hallucinated” case citation	“In the Court’s view, there is nothing inherently wrong with an attorney properly and competently utilizing AI or any of its subsets to practice law or litigate cases. [But a] basic prerequisite to the filing of any … paper in court is for the drafting and filing attorney(s) to carefully check every case citation, fact, and argument to make sure that they are correct and proper. Attorneys cannot delegate that role to AI, computers, robots, or any other form of technology.”	Pay counsel’s fees. Complete CLE on AI. $1,000 and $500 fine.	Lawyer
Alharoun v Qatar National Bank and QNB	UK	May 25	TBC	TBC	“In CL-2024-000435, it appears from the Order of Mrs Justice Dias that correspondence was sent to the court, and witness statements were filed, citing authorities that do not exist and claiming that other authorities contained passages that they do not contain” Rt Hon. Dame Victoria Sharp	TBC – Not clear if AI but linked with Ayinde v Haringey below so keep under review.	TBC	Link
R (Ayinde) v Haringey	UK	May 25	AI use disputed.	Citation of Fake Cases	“ It is such a professional shame. The submission was a good one. The medical evidence was strong. The ground was potentially good. Why put a fake case in?” “I should say it is the responsibility of the legal team, including the solicitors, to see that the statement of facts and grounds are correct.” “…I consider that it would have been negligent for this barrister, if she used AI and did not check it, to put that text into her pleading.” Mr Justice Richie	Wasted costs £2,000 Counsel and £2,000 Solicitor and referral to regulators. FURTHER HEARING 23 May	Lawyers (accused not found)	Link Link
Willis v Bank National Association	US	May 25	N/A	Court discusses pros and cons of AI and allows careful use.	“[b]y presenting to the court a pleading, written motion, or other paper – whether by signing, filing, submitting, or later advocating it – an attorney or unrepresented party certifies that to the best of the person’s knowledge, information, and belief, formed after an inquiry reasonable under the circumstances ․ the claims, defenses, and other legal contentions are warranted by existing law…” “…[c]onfirming a case is good law is a basic, routine matter and something to be expected from a practicing attorney…”	Court issued standing order – rules in relation to AI use	N/A	link
Concord MG v Anthropic PBC	USA	May 25	Claud (Anthropic)	Allegation cited article was invented by Claude	“The Court gave Anthropic time to investigate the circumstances surrounding the challenged citation. …the Court finds this issue is a serious one—if not quite so grave as it at first appeared. Anthropic’s counsel protests that this was “an honest citation mistake” but admits that Claude.ai was used to “properly format” at least three citations and, in doing so, generated a fictitious article name with inaccurate authors (who have never worked together) for the citation at issue… That is a plain and simple AI hallucination. Yet the underlying article exists, was properly linked to and was located by a human being using Google search; so, this is not a case where “attorneys and experts [have] abdicate[d] their independent judgment and critical thinking skills in favor of ready-made, AI-generated answers….” “…A remaining serious concern, however, is Anthropic’s attestation that a “manual citation check” was performed but “did not catch th[e] error.”	Written explanation Ordered	Expert/Lawyer	Link
Nexgen Pathology Services Ltd v Darceuil Duncan	Trinidad & Tobago	May 25	Unverified Google/ChatGPT research (admitted “online source now unavailable”)	non‑existent cases.	69. The Court acknowledges that digital tools including AI and internet-based platforms, are increasingly common and valuable in legal research; indeed, this Court itself makes use of such tools where appropriate. However, their use must be accompanied by discernment and subjected to rigorous verification. This is because AI-generated content is susceptible to producing what are commonly referred to as “hallucinations”: fabricated, yet plausible-sounding outputs that may result from gaps or limitations in the model’s underlying data. Legal practitioners must not rely on such tools uncritically. Any information obtained through these means must be independently verified before being presented to the Court. 70. The Court emphasizes that citing non-existent cases, even inadvertently, constitutes a serious abuse of process and professionalism. It risks misleading the Court, prejudicing the opposing party, and eroding public confidence in the administration of justice. Counsel are reminded that the duty of candour to the Court requires that they verify the authenticity of every case cited. If any material has been generated with the assistance of AI or other non-traditional sources, full disclosure to the Court is both appropriate and expected.” The Hon. Mr. Justice Westmin R.A. James	condemned the “irresponsible use of generative AI,” struck the authorities, referred both counsel to the Law Association’s Disciplinary Committee, and reminded practitioners of their duty to verify AI‑derived material	Lawyer	Link
Ko v Li	Canada	May 25	ChatGPT (suspected)	multiple non-existent or mis-linked authorities	[14] This occurrence seems similar to cases in which people have had factums drafted by generative artificial intelligence applications (like ChatGPT). Some of these applications have been found to sometimes create fake legal citations that have been dubbed “hallucinations.” It appears that Ms. Lee’s factum may have been created by AI and that before filing the factum and relying on it in court, she might not have checked to make sure the cases were real or supported the propositions of law which she submitted to the court in writing and then again orally. FL Myers J	show cause why should not be cited for contempt.	Lawyer	Link
Reclamação 78.890	Brazil	May 25	TBC – awaiting verified English translation of the judgment.	TBC – awaiting verified English translation of the judgment.	TBC – awaiting verified English translation of the judgment.	Summary withheld until a reliable translation is available. Readers should consult the original Portuguese decision linked in the final column.	TBC	Link
Lacey v State Farm General Ins	US	May 25	Google Gemini + Westlaw Precision (CoCounsel)	approximately nine of the 27 legal citations in the ten-page brief were incorrect in some way. At least two of the authorities cited do not exist at all. Additionally, several quotations attributed to the cited judicial opinions were phony and did not accurately represent those materials.		supplemental briefs struck, no further discovery relief, Lawyers to pay compensation $31,100	Lawyer	Link
Ramirez v Humala	US	May 25	Uncertain	Reply letter cited four non-existent cases.		$1,000 monetary sanction and service of order on client	Lawyer	Link
Bandla v SRA	UK	May 25	Google Searches, not Gen-AI	fake or mis-described case authorities in appeal grounds and skeleton.		Citation caused grounds of appeal to be struck out as abuse of process.	Litigant in Person (former solicitor)	Link
ZZaman v Revenue & Customs	UK	Apr 25	Unspecified	False Cases	29. However, our conclusion was that Mr Zzaman’s statement of case, written with the assistance of AI, did not provide grounds for allowing his appeal. Although some of the case citations in Mr Zzaman’s statement were inaccurate, the use of AI did not appear to have led to the citing of fictitious cases (in contrast to what had happened in Felicity Harber v HMRC [2023] UKFTT 1007 (TC) ). But our conclusion was that the cases cited did not provide authority for the propositions that were advanced. This highlights the dangers of reliance on AI tools without human checks to confirm that assertions the tool is generating are accurate. Litigants using AI tools for legal research would be well advised to check carefully what it produces and any authorities that are referenced. These tools may not have access to the authorities required to produce an accurate answer, may not fully “understand” what is being asked or may miss relevant materials. When this happens, AI tools may produce an answer that seems plausible, but which is not accurate. These tools may create fake authorities (as seemed to be the case in Harber ) or use the names of cases to which it does have access but which are not relevant to the answer being sought (as was the case in this appeal). There is no reliable way to stop this, but the dangers can be reduced by the use of clear prompts, asking the tool to cite specific paragraphs of authorities (so that it is easy to check if the paragraphs support the argument advanced), checking to see the tool has access to live internet data, asking the tool not to provide an answer if it is not sure and asking the tool for information on the shortcomings of the case being advanced. Otherwise there is a significant danger that the use of an AI tool may lead to material being put before the court that serves no one well, since it raises the expectations of litigants and wastes the court’s time and that of opposing parties.	Guidance of how to avoid issues.	Litigant in Person	link
Saxena v Martínez-Hernández 2025 WL 1194003	US	Apr 25	Unspecified	AI hallucinations that could not be located.		Warning that Gen AI citations may trigger sanctions	Litigant in Person	link
Bevins v Colgate-Palmolive Co	US	Apr 25	Unspecified	Both citations appear to be artificial intelligence (“AI”) “hallucinations” made up of parts of actual cases [read full footnote 10]		serve order on Bar, appearance stricken, inform and new counsel.	Lawyer	Link Link
Coomer v Lindell et al (My Pillow)	US	Apr 25	Unspecified	the Court identified nearly thirty defective citations		Order to Show Cause directing counsel and firm to explain why they, their clients and in-house lawyers should not be sanctioned and referred to disciplinary authorities, and to certify personal service of the order.	Lawyer	Link
Dehghani v. Castro	US	Apr 25	ChatGPT (likely the handiwork of a ChatGPT-style AI program’s hallucination)	Six non-existent authorities		$1,500 fine, 1-hour CLE AI/Ethics and report to Bar	Lawyer	Link
Benjamin v Costco Wholesale Corp	US	Apr 25	ChatOn (mobile generative-AI app)	Reply in support of a motion to remand cited four non-existent cases created by ChatOn		$1 000 fine, ordered counsel to serve the order on her client and file proof of service,	Lawyer	Link
Williams v Capital One	US	Mar 25	CoCounsel	Non existent cases – Judge named the false citations in judgment	“It is not acceptable for parties to submit “filings to the Court containing citations to legal authority that does not exist, whether drafted with the assistance of artificial intelligence or not. “	Dismissed and warning.	Lawyer	link
Nguyen v Savage Enterprises	US	Mar 25	Uncertain (AI… may have crept in)	citing nonexistent authority in brief		show cause	Lawyer	Link
Mid Central Operating Engineers	US	Feb 25	Unspecified	Three fake case citations.		$15,000 in sanctions (US $5 000 per brief), referral for further professional discipline, and an order that counsel notify.	Lawyer	Link
Unnamed appeal (TJ Santa Catarina, Boletim-TJSC	Brazil	Feb 25	TBC – awaiting verified English translation of the judgment.	TBC – awaiting verified English translation of the judgment.		Summary withheld until a reliable translation is available. Readers should consult the original Portuguese decision linked in the final column.	TBC	Link
Wadsworth v Walmart Inc	US	Feb 25	MX2.law	Eight non-existent authorities		Main counsel’s pro hac vice admission revoked; barred from case; fined US $3,000, plus US $1,000 fines for certain other counsel.	Lawyer	Link
Bunce v Visual Tech. Innovations, Inc	US	Feb 25	ChatGPT	non-existent and mis-characterised cases		Show cause, US $2 500 penalty, completion of a one-hour CLE on AI & ethics, and warned further breaches would draw harsher pen	Lawyer	Link
Valu v Minister for Immigration	Aus	Jan 25	Unspecified	Fabricated citations		lawyer referred to regulator and caution given about AI	Lawyer	link
Olsen v Finansiel Stabilitet	UK	Jan 25	Unspecified	The appellants submitted a summary of a non-existent case, which was included in their authorities bundle.	“I have narrowly and somewhat reluctantly come to the conclusion that I should not cause a summons for contempt of court to be issued to the appellants under CPR rule 81.6. I do not think it likely that a judge (whether myself or another judge) could be sure, to the criminal standard of proof, that the appellants knew the case summary was a fake. They may have known but they could not be compelled to answer questions about the identity of the person who supplied it.” Mr Justice Kerr	While considering contempt proceedings, the judge ultimately decided against it, noting the appellants’ lack of legal representation and cooperation.	Litigant in Person	link
Mavundla v MEC	South Afrrica	Jan 25	ChatGPT/Meta (suspected)	cited multiple non-existent or mis-quoted authorities.		Referred to Legal Practice Council for investigation and further action. Attorneys pay costs of certain appearances.	Lawyer	Link
Kohls and Franson v Ellison	US	Jan 25	GPT-4o	Two non-existent academic articles and mis-attributed a third.		Expert witness evidence excluded.	Expert Witness	Link
United States v Hayes	US	Jan 25	Unspecified	Discussion about citation hallucination and citation error		Show cause.	Lawyer	Link
Gauthier v Goodyear Tire & Rubber	US	Nov 24	Claud (Anthropic)	summary-judgment response quoted two imaginary cases and several fabricated quotations generated with Claude; after a show-cause order, counsel admitted the error		$2 000 penalty, required completion of a 1-hour CLE on AI/ethics, and ordered service of the order on the client	Lawyer	Link
Handa v Mallick	Aus	Jul 24	LEAP	generated list of authorities		practitioner afforded an opportunity make submissions as to why conduct in tendering the list of authorities should not be referred	Lawyer	link
Grant v City of Long Beach 96 F.4th 1255	US	Mar 24	Unspecified	filed an opening brief replete with misrepresentations and fabricated case law. The brief included only a handful of accurate citations, almost all of which were of little use to this Court because they were not accompanied by coherent explanations of how they supported appellants’ claims.		struck brief and dismissed appeal.	Lawyer	link
Zhang v Chen	Canada	Feb 24	ChatGPT	Citation of two non-existent cases.	Masuhara J. held: “…Citing fake cases in court filings and other materials handed up to the court is an abuse of process and is tantamount to making a false statement to the court. Unchecked, it can lead to a miscarriage of justice…”	Lawyer pay costs personally and review files	Lawyer	Link
Park v Kim	US	Jan 24	ChatGPT	Single non-existent state-court decision cited in the appellant’s reply brief		Lawyer referred to grievance panel and serve order on client.	Lawyer	Link
Vanguard Construction & Development Co. v 400 Times Square Associates, LLC (2025).	US	24	ChatGPT (alleged)	the brief was accused of being AI-generated, but the court found the insinuation unsubstantiated		Court admonished counsel for making an evidence-free allegation; no sanctions were imposed	Lawyers (accused not found)	Link
Crypto Open Patent Alliance v Dr. Craig Steven Wright	UK	24	ChatGPT (likely not certain)	Series of authorities that do not contain passages attributed.		false references unlikely to be deliberate. Principles not in doubt so court didn’t engage further.	Litigant in Person	Link
Sala Primera del Tribunal Constitucional – Nota Informativa 90/2024	Spain	TBC	TBC – awaiting verified English translation of the judgment.	TBC – awaiting verified English translation of the judgment.		Summary withheld until a reliable translation is available. Readers should consult the original Portuguese decision linked in the final column.	TBC	Link
Al-Hamim v Star Hearthstone 2024COA128	US	Dec 24	Unspecified	Fake authorities		Declined sanctions but warned future penalties for AI misuse	Litigant in Person	link
Mortazavi v Hamilton	US	Sep 24	Uncertain	Motion to remand contained a citation to a non-existent case and other errors produced with AI; declaration (no AI disc)		Order to Show Cause	Lawyer	Link
Dayal	Aus	Aug 24	Unnamed AI in practice-management suite	Non existent cases		Referred to professional body	Lawyer	link
US v Cohen	US	Mar 24	Google Bard	Non-existent cases		Court declined to impose sanctions for the citation in that motion to nonexistent cases.	Lawyer	“` :contentReference[oaicite:0]{index=0}
Kruse v Karlen	US	Feb 24	Unspecified generative-AI via hired online “consultant” purporting to be an attorney to prepare the Brief. Indicated that the fee paid amounted to less than one percent of the cost of retaining an attorney.	Apologised for submitting fictitious cases. Did not know. denied intention to mislead.		court branded the appeal frivolous, dismissed it and damages.	Litigant in Person	Link
Moffatt v Air Canada	Canada	Feb 24	website customer-service chatbot	Chatbot suggested passenger could apply for bereavement fares retroactively. Passenger later learned airline did not permit retroactive applications.		Negligent Misrep. $650.88 damages, CA $36.14 interest, and CA $125 tribunal fees.	Chatbot	Link
Smith v Farwell	US	Feb 24	Unspecified	Fabricated US Supreme Court cases		$2,000 sanction on counsel and warning to bar.	Lawyer	link
re Thomas G. Neusom	US	Jan 24	Unspecified	inaccurate citations and fabricated authorities		Suspension and corrective steps.	Laywer	Link Link
Will of Samuel	US	Jan 24	Unspecified	Multiple fictional citations		Further hearing set.	Lawyer	Link
Harber v HMRC	UK	Dec 23	ChatGPT	Nine fabricated FTT decisions.		Tribunal criticised the wasted time and money caused.	Litigant in Person	Link
Thomas v Pangburn 2023 WL 9425765	US	Oct 23	Unspecified	misleading citations		Court highlighted dangers	Litigant in Person	link
People v Crabill	US	Nov 23	ChatGPT	Cited non-existent cases and mis-described cases.		one year and one day suspension, with ninety days to be served and the remainder to be stayed upon successful completion of a two year period of probation, with conditions.	Lawyer	Link
Ex parte Lee	US	Jul 23	Unspecified	False case citation		court noted brief “may have been prepared by AI” but imposed no sanctions; stressed verification duty	Lawyer	link
Parker v Forsyth N.O.	South Africa	Jun 23	ChatGPT	Cases cited – names and citations are fictitious, the facts are fictitious, and the decisions are fictitious.	“The Plaintiff’s attorneys used this artificial intelligence medium to conduct legal research and accepted the results that it generated without satisfying themselves as to its accuracy. As it turned out, the cases listed above do not exist. The names and citations are fictitious, the facts are fictitious, and the decisions are fictitious. The Plaintiff’s counsel was constrained to concede as much. ” [87] A Chaitram Regional Magistrate	Judge held the lawyers had shown “undue faith in AI” and, although not wilful, were “over-zealous and careless.” “The embarrassment associated with this incident is probably sufficient punishment for the Plaintiff’s attorneys.”	Lawyer	Link
Scott v Federal National Mortgage Ass’n (Me. Supe	US	Jun 23	Unspecified	fictitious authorities		Complaint dismissed and Rule 11 sanctions ordered	Litigant in Person	link

Generative AI is evolving rapidly, and so are the legal and ethical standards surrounding its use. By following this tracker you can see how courts across the globe respond to fabricated citations, ranging from mild responses to hefty sanctions. Please keep revisiting. New decisions arrive often, reinforcing the importance of verifying sources and practising diligent, technology-aware advocacy in every jurisdiction, daily.

AI Hallucination Cases Tracker FAQ