Skip to content
Founderz: Responsible Use of AI home

in collaboration with Microsoft

All case studies

Case study · Court decisions

Invented case law in court filings

Lawyers were fined for citing cases an AI chatbot invented. A public database now lists more than 2,000 similar court and tribunal decisions.

Principles involved

When and where

2023 to 2026

Worldwide, mostly the United States

What happened

In 2023, two New York lawyers filed a brief in Mata v. Avianca that cited court decisions an AI chatbot had invented. When the court questioned whether the cases existed, they kept defending them. On June 22, 2023, the judge found they had acted in bad faith, fined them and their firm US$5,000, and ordered them to write to their client and to each judge falsely named as the author of a fake opinion.

It became the best-known early case, and many more followed. By September 23, 2026, a public database maintained by the researcher Damien Charlotin listed 2,077 decisions in which a court or tribunal found, or implied, that a party had relied on AI-invented content, mostly fake citations. Of those, 1,428 were in the United States.

What the sources establish

The sanctions order establishes the facts of Mata v. Avianca. The database counts decisions in which courts found or implied reliance on AI-invented content. It leaves out cases where this was only alleged, and one person maintains it, so treat the count as a minimum.

Implications

Professionals with every reason to be careful still filed invented cases, because the output looked exactly like real legal writing. Specialist tools help without solving it: a peer-reviewed Stanford study found that leading AI legal research tools hallucinated on 17% to 33% of test queries. So open each source, however polished the text looks.

Recommended practice

  1. 01Open every citation before it goes into anything official.
  2. 02Be most careful where the output looks most authoritative: legal, financial and technical documents.
  3. 03If you find an invented source in something you sent, correct the record quickly and say so.

Sources

What each source establishes, and its limits. The practices and recommendations on this page are ours, and the facts come from the sources. See every source we use.

  1. Mata v. Avianca, Inc.: opinion and order on sanctions US District Court, Southern District of New York · June 22, 2023 · Court decision Found that lawyers acted in bad faith by filing, and then defending, court opinions invented by an AI chatbot; fined them and their firm US$5,000 and ordered letters to their client and to the judges falsely named. Limits: One case. Its lessons are about professional duties more than the technology.
  2. AI Hallucination Cases database Damien Charlotin · Updated continuously (2,077 cases on September 23, 2026) · Research database Lists decisions in which a court or tribunal found, or implied, that a party relied on AI-invented content, mostly fake citations: 2,077 by September 23, 2026, 1,428 of them in the United States. Limits: Maintained by one person. It leaves out cases where reliance was only alleged.
  3. Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools Journal of Empirical Legal Studies (Stanford RegLab and HAI researchers) · April 23, 2025 · Peer-reviewed research On 202 legal questions, leading AI legal research tools hallucinated on 17% to 33% of queries, and a general-purpose model on 43%. Limits: A snapshot of tools tested in 2024. The tools have changed since.

Cite this page

Founderz (2026). Invented case law in court filings. The RUAI Standard, 2026 edition. Developed by Founderz in collaboration with Microsoft. https://responsibleai.founderz.com/toolkit/case-studies/invented-case-law-in-court-filings

Licensed under CC BY 4.0: share and adapt with attribution.