Within Hallucinations
When AI invents sources that look real
The Mata v.
On this page
- What happened in Mata v. Avianca
- Why fabricated citations are hard to spot
- Verification habits that prevent source errors
Page outline Jump by section
Introduction
The 2023 case Mata v. Avianca became one of the most widely discussed examples of artificial intelligence producing information that looked authoritative but was entirely false. In a court filing, lawyers cited legal cases that did not exist. The citations had been generated by ChatGPT, and the attorneys submitted them without independently verifying that the cases were real. When the court investigated, it discovered that the cited authorities contained fabricated case names, invented quotations, and nonexistent legal reasoning. The resulting sanctions turned the case into a landmark warning about AI hallucinations and the cost of trusting plausible text without checking the underlying evidence. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
For anyone trying to understand artificial intelligence, Mata v. Avianca illustrates a crucial lesson: a language model can generate references that look exactly like genuine sources even when no such sources exist. The problem was not poor grammar or obvious nonsense. The problem was credibility without verification. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
What happened in Mata v. Avianca
The underlying lawsuit concerned a personal injury claim by Roberto Mata against Avianca, an airline. During the litigation, Avianca sought dismissal of the case. In response, lawyers representing Mata submitted a brief that relied on several judicial decisions as supporting authority. [CaseMine]casemine.comCase Mine Mata vAvianca, Inc. | 22-cv-1461 (PKC) | S.D.N.Y.Jun 22, 2023 — Plaintiff Roberto Mata asserts that on August 27 or 28, 2019, he was severely i…
The difficulty emerged when Avianca’s lawyers and the court attempted to locate some of the cited cases. They could not find them in legal databases. The court then ordered the attorneys to provide copies of the authorities they had cited. Instead of resolving the issue, the submitted materials deepened concerns because the supposed opinions contained further inconsistencies and signs of fabrication. [Justia Law]law.justia.comLaw Mata vAvianca, Inc., No. 1:2022cv01461 - Document 54…22 Jun 2023 —) A fake opinion is not “existing law” and citation to a fake opinion doe…
Court records showed that attorney Steven Schwartz had used ChatGPT to help conduct legal research. According to the sanctions opinion, the chatbot generated multiple cases that appeared genuine but were fictitious. Some included detailed summaries, quotations, citations, and procedural histories that had never existed. ChatGPT reportedly reassured the attorney that the cases were real and could be found in recognised legal databases, even though they could not. [Wikipedia]WikipediaMata v. Avianca, IncMata v. Avianca, Inc
Judge P. Kevin Castel described the situation as involving “fake judicial opinions” with fake quotations and citations generated by artificial intelligence. The court concluded that the lawyers had failed in their professional duty to verify the authorities they presented. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
The sanctions
In June 2023, Judge Castel imposed sanctions on the lawyers involved and their law firm. The court ordered them to pay a total penalty of $5,000 and required additional corrective measures. The opinion stressed that lawyers remain responsible for the accuracy of their filings regardless of whether AI tools are used during research or drafting. [Reuters+2The Time Blawg]reuters.comnew york lawyers sanctioned using fake chatgpt cases legal brief 2023 06 22New York lawyers sanctioned for using fake ChatGPT…26 Jun 2023 — A U.S. judge on Thursday imposed sanctions on two New York law…
A key point in the ruling was that the court did not condemn AI use itself. Instead, it criticised the failure to verify AI-generated material before presenting it as legal authority. Judge Castel wrote that existing professional obligations continue to apply even when attorneys use new technologies. [LawSites]lawnext.comLawSitesCourt Imposes Sanctions On Lawyers Who Filed Bogus…22 Jun 2023 — Kevin Castel, “existing rules impose a gatekeeping role on at…
Why fabricated citations are hard to spot
The Mata incident became famous because the fake citations did not look fake.
The invented cases followed familiar legal conventions. They had plausible party names, realistic formatting, references to legal principles, and apparently coherent reasoning. To someone who did not independently search for the cases, the citations could appear authentic. [Wikipedia]WikipediaMata v. Avianca, IncMata v. Avianca, Inc
This reflects a broader characteristic of generative AI systems. Large language models are trained to predict convincing sequences of words. They can produce text that resembles genuine legal writing, academic references, or news reporting without actually retrieving or verifying the underlying facts. As a result, fabricated information may be presented with the same confidence and structure as accurate information. [internationaltaxjournal.online]internationaltaxjournal.onlineAvianca exposed this risk when attorneys relied on ChatGPT-generated research and submitted fictitious case law, resulting in…Read more…
Several factors make these hallucinated citations especially deceptive:
- Realistic formatting: The citations resemble authentic legal references.
- Detailed specificity: Invented cases often include dates, quotations, procedural histories, and legal doctrines.
- Internal consistency: The fabricated material may fit logically with the surrounding argument.
- False reassurance: When questioned, some AI systems may continue generating confident explanations rather than admitting uncertainty. [Law Society Journal]lsj.com.aulawyer faces sanctions for using bogus citations from chatgptLaw Society JournalLawyer faces sanctions for using 'bogus' citations from…29 May 2023 — Schwartz also asked the chatbot if any of the…
The Mata filings demonstrated all of these features. The problem was not a lack of detail; it was that the detail itself had been invented. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
Why this case became a turning point
Many examples of AI hallucinations involve incorrect trivia, mistaken facts, or fabricated quotations. Mata v. Avianca was different because the hallucinations entered a formal legal proceeding, where citations are expected to be verifiable evidence rather than suggestions. [Reuters]reuters.comnew york lawyers sanctioned using fake chatgpt cases legal brief 2023 06 22New York lawyers sanctioned for using fake ChatGPT…26 Jun 2023 — A U.S. judge on Thursday imposed sanctions on two New York law…
The case received extensive attention from courts, law firms, legal educators, and professional organisations because it transformed an abstract concern into a concrete demonstration of risk. Before Mata, warnings about AI-generated misinformation were often theoretical. After Mata, there was a documented example showing how fabricated sources could pass through professional workflows and reach a federal court filing. [Association of Corporate Counsel (ACC)]acc.com8 Aug 2023 — The recent story of two New York attorneys “duped” by ChatGPT into citing “fake” cases in a court submission— and the sancti…
The incident has since been cited in discussions about legal ethics, AI governance, professional responsibility, and the need for human oversight when using generative AI tools. It is frequently referenced as a cautionary case study in training materials for lawyers and other professionals who rely on evidence-based work. [jdsupra.com+2SSRN]jdsupra.comfederal court turns up the heat on 1849454Avianca court imposed monetary fines and other remedial measures against the offending attorneys, citing Rule 11 of the Federal…Read more…
Verification habits that prevent source errors
The most important lesson from Mata v. Avianca is not that AI should never be used. It is that AI-generated references should never be treated as verified evidence simply because they look convincing. [LawSites]lawnext.comLawSitesCourt Imposes Sanctions On Lawyers Who Filed Bogus…22 Jun 2023 — Kevin Castel, “existing rules impose a gatekeeping role on at…
Several practical habits can reduce the risk of fabricated citations:
Check every cited source directly
A citation should be located in the original database, publication, or court record before being relied upon. If a source cannot be found independently, it should not be treated as evidence. [Association of Corporate Counsel (ACC)]acc.com8 Aug 2023 — The recent story of two New York attorneys “duped” by ChatGPT into citing “fake” cases in a court submission— and the sancti…
Treat AI output as a draft, not an authority
Generative AI can help identify topics, suggest search terms, or summarise material. However, the system itself is not proof that a source exists. Every factual claim and citation requires separate confirmation. [internationaltaxjournal.online]internationaltaxjournal.onlineAvianca exposed this risk when attorneys relied on ChatGPT-generated research and submitted fictitious case law, resulting in…Read more…
Be cautious when details seem unusually perfect
Hallucinated citations often contain rich detail. Ironically, specificity can increase confidence even when the information is false. Verification should become more rigorous, not less, when a source appears highly polished. [Goldberg Segalla]goldbergsegalla.comGPT's ability to lie and create false information and argued that there was no bad faith. A hearing was held on June 8, 2023.Read more…
Verify quotations as well as citations
A source may exist while a quotation does not. Good verification involves checking the original document rather than relying solely on an AI-generated summary. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
What Mata v. Avianca reveals about AI
The enduring significance of Mata v. Avianca lies in how clearly it exposed a central limitation of generative AI. A language model can produce text that resembles evidence without actually being evidence. It can generate case names, quotations, and citations that satisfy the patterns of legal writing while lacking any connection to real-world sources. [UC Berkeley Law]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
The case therefore serves as a powerful illustration of the broader challenge of AI hallucinations. Fluent language can create an impression of reliability, but reliability comes from verification. In Mata v. Avianca, the fabricated citations looked authentic enough to reach a federal court. The consequences demonstrated why checking sources remains a human responsibility, even when AI systems make the text appear complete and convincing. [UC Berkeley Law+2Reuters]law.berkeley.eduUC Berkeley Law Mata vAvianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence…
eBay marketplace picks
Marketplace Samples
Example marketplace items related to this page. Use the search link to explore similar finds on eBay.
Endnotes
-
Source: law.berkeley.edu
Title: UC Berkeley Law Mata v
Link: https://www.law.berkeley.edu/wp-content/uploads/archive/2025/12/Mata-v-Avianca-Inc.pdfSource snippet
Avianca, Inc., 678 F.Supp.3d 443 (2023)existent judicial opinions with fake quotes and fake citations created by artificial intelligence...
-
Source: reuters.com
Title: new york lawyers sanctioned using fake chatgpt cases legal brief 2023 06 22
Link: https://www.reuters.com/legal/new-york-lawyers-sanctioned-using-fake-chatgpt-cases-legal-brief-2023-06-22/Source snippet
New York lawyers sanctioned for using fake ChatGPT...26 Jun 2023 — A U.S. judge on Thursday imposed sanctions on two New York law...
-
Source: acc.com
Link: https://www.acc.com/resource-library/practical-lessons-attorney-ai-missteps-mata-v-aviancaSource snippet
8 Aug 2023 — The recent story of two New York attorneys “duped” by ChatGPT into citing “fake” cases in a court submission— and the sancti...
-
Source: casemine.com
Title: Case Mine Mata v
Link: https://www.casemine.com/judgement/us/6499235932898b72d5cfb1aaSource snippet
Avianca, Inc. | 22-cv-1461 (PKC) | S.D.N.Y.Jun 22, 2023 — Plaintiff Roberto Mata asserts that on August 27 or 28, 2019, he was severely i...
-
Source: law.justia.com
Title: Law Mata v
Link: https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1%3A2022cv01461/575368/54/Source snippet
Avianca, Inc., No. 1:2022cv01461 - Document 54...22 Jun 2023 —) A fake opinion is not “existing law” and citation to a fake opinion doe...
-
Source: Wikipedia
Title: Mata v. Avianca, Inc
Link: https://en.wikipedia.org/wiki/Mata_v._Avianca%2C_Inc -
Source: internationaltaxjournal.online
Link: https://internationaltaxjournal.online/index.php/itj/article/view/505Source snippet
Avianca exposed this risk when attorneys relied on ChatGPT-generated research and submitted fictitious case law, resulting in...Read more...
-
Source: papers.ssrn.com
Link: https://papers.ssrn.com/sol3/Delivery.cfm/5005967.pdf?abstractid=5005967&mirid=1Source snippet
AI, Fake Law and Professional Guidance31 Oct 2024 — Given the publicity surrounding Mata v Avianca, one might expect lawyers to be cautio...
-
Source: jdsupra.com
Title: federal court turns up the heat on 1849454
Link: https://www.jdsupra.com/legalnews/federal-court-turns-up-the-heat-on-1849454/Source snippet
Avianca court imposed monetary fines and other remedial measures against the offending attorneys, citing Rule 11 of the Federal...Read more...
-
Source: Wikipedia
Title: Juan Mata
Link: https://en.wikipedia.org/wiki/Juan_MataSource snippet
Juan MataJuan Manuel Mata García (Spanish: [xwan maˈnwel ˈmata ɡaɾˈθi.a]; born 28 April 1988) is a Spanish professional footballer wh...
Published: April 1988
-
Source: law.justia.com
Link: https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1%3A2022cv01461/575368/55/Source snippet
Avianca, Inc., No. 1:2022cv01461 - Document 55...Avianca's motion to dismiss is GRANTED. (ECF 16.) The Clerk is respectfully directed to...
-
Source: papers.ssrn.com
Link: https://papers.ssrn.com/sol3/Delivery.cfm/5005967.pdf?abstractid=5005967&mirid=1&type=2Source snippet
AI, Fake Law and Professional Guidance31 Oct 2024 — Given the publicity surrounding Mata v Avianca, one might expect lawyers to be cautio...
-
Source: reuters.com
Title: lawyers face sanctions citing fake cases with ai warns uk judge 2025 06 06
Link: https://www.reuters.com/world/uk/lawyers-face-sanctions-citing-fake-cases-with-ai-warns-uk-judge-2025-06-06/Source snippet
Lawyers face sanctions for citing fake cases with AI, warns...6 Jun 2025 — Lawyers who use artificial intelligence to cite non-existent...
-
Source: proskauer.com
Link: https://www.proskauer.com/blog/artificially-unintelligent-attorneys-sanctioned-for-misuse-of-chatgptSource snippet
Artificially Unintelligent: Attorneys Sanctioned for Misuse of...30 Jun 2023 — Mata's lawyers then filed a 10-page brief, the subject of...
-
Source: lsj.com.au
Title: lawyer faces sanctions for using bogus citations from chatgpt
Link: https://lsj.com.au/articles/lawyer-faces-sanctions-for-using-bogus-citations-from-chatgpt/Source snippet
Law Society JournalLawyer faces sanctions for using 'bogus' citations from...29 May 2023 — Schwartz also asked the chatbot if any of the...
Published: May 2023
-
Source: lawnext.com
Link: https://www.lawnext.com/2023/06/court-imposes-sanctions-on-lawyers-who-filed-bogus-cases-after-relying-on-chatgpt-for-legal-research.htmlSource snippet
LawSitesCourt Imposes Sanctions On Lawyers Who Filed Bogus...22 Jun 2023 — Kevin Castel, “existing rules impose a gatekeeping role on at...
-
Source: thetimeblawg.com
Title: The Time Blawg Chat GPT Lawyers Sanctioned
Link: https://thetimeblawg.com/2023/06/24/chatgpt-lawyers-sanctioned/Source snippet
Kevin Castel issued a 34 page Opinion [PDF]. It concludes that a penalty of $5,000 is jointly and severally imposed on the lawyers in...
-
Source: goldbergsegalla.com
Link: https://www.goldbergsegalla.com/app/uploads/2023/10/Fake-Cases-Real-Consequences-Misuse-of-ChatGPT-Christoper-F.-Lyon-NY-Litigator.pdfSource snippet
GPT's ability to lie and create false information and argued that there was no bad faith. A hearing was held on June 8, 2023.Read more...
Published: June 8, 2023
-
Source: quimbee.com
Title: Mata v
Link: https://www.quimbee.com/cases/mata-v-avianca-inc-678-f-supp-3d-443-2023Source snippet
Avianca, Inc., 678 F. Supp. 3d 443 (2023)Roberto Mata (plaintiff) sued Avianca, Inc. (defendant). Avianca filed a motion to dismiss. Mata...
Additional References
-
Source: medium.com
Link: https://medium.com/%40sheldonksalmon/ai-incident-analysis-report-lawyer-sanctioned-for-submitting-ai-generated-fake-citations-mata-v-7f68d66996d5Source snippet
AI Incident Analysis Report — Lawyer Sanctioned for...WARNING: Attorneys sanctioned for unverified AI citations (Mata v. Avianca). Verif...
-
Source: linkedin.com
Link: https://www.linkedin.com/pulse/mata-v-avianca-blame-solely-chatgpt-robert-rothSource snippet
Mata v. Avianca: The Blame is not Solely with ChatGPTJudge Kevin Castel was having none of it. He meticulously questioned each lawyer. Fi...
-
Source: seyfarth.com
Title: update on the chatgpt case counsel who submitted fake cases are sanctioned
Link: https://www.seyfarth.com/news-insights/update-on-the-chatgpt-case-counsel-who-submitted-fake-cases-are-sanctioned.htmlSource snippet
Update on the ChatGPT Case: Counsel Who Submitted...26 Jun 2023 — On June 8, 2023, the judge in the Mata case held a hearing on the issu...
Published: June 8, 2023
-
Source: northerntrust.com
Link: https://www.northerntrust.com/content/dam/northerntrust/events/documents/look-before-you-leap.pdfSource snippet
ChatGPT... Measures for the Management of Generative Artificial Intelligence Services (Draft...Read more...
-
Source: virtuositylegal.com
Title: ai in court when legal tech goes rogue lessons from mata v avianca
Link: https://virtuositylegal.com/ai-in-court-when-legal-tech-goes-rogue-lessons-from-mata-v-avianca/Source snippet
AI in Court: When Legal Tech Goes Rogue30 Apr 2025 — These fake cases were generated by ChatGPT, an AI-powered language model, which fals...
-
Source: esquiresolutions.com
Title: federal court turns up the heat on attorneys using chatgpt for research
Link: https://www.esquiresolutions.com/federal-court-turns-up-the-heat-on-attorneys-using-chatgpt-for-research/Source snippet
Avianca court imposed monetary fines and other remedial measures against the offending attorneys, citing Rule 11 of the Federal...
-
Source: hinshawlaw.com
Title: Hinshaw & Culbertson LLPNew York Lawyer Who Cited Non-Existent Cases
Link: https://www.hinshawlaw.com/en/insights/lawyers-for-the-profession-alert/new-york-lawyer-who-cited-non-existent-cases-generated-by-chatgpt-now-faces-sanctions-hearingSource snippet
Kevin Castel) has issued an order to show cause against a plaintiff's attorney who had cited non-existent or “bogus” cases he...
-
Source: medium.com
Title: When AI Goes Wrong in the Courtroom | by Mrinal Kanti
Link: https://medium.com/%40mrinal.k.sardar/when-ai-goes-wrong-in-the-courtroom-c532275ccf6cSource snippet
judicial opinions, fake quotations, and bogus internal citations.... Avianca, artificial intelligence continues to transform legal pract...
-
Source: forbes.com
Title: Lawyer Used Chat GPT In Court—And Cited Fake Cases
Link: https://www.forbes.com/sites/mollybohannon/2023/06/08/lawyer-used-chatgpt-in-court-and-cited-fake-cases-a-judge-is-considering-sanctions/Source snippet
A...8 Jun 2023 — A member of Mata's law team then revealed he had used ChatGPT to conduct legal research for the court filing that refer...
-
Source: documentcloud.org
Link: https://www.documentcloud.org/documents/23856993-sanctions-mata-v-avianca/Source snippet
File... Search. Search... p. 1... Share... p. 2... Share... p. 3... Share... p. 4... Share... p. 5... Share... p. 6... Share...
Topic Tree