Courts on AI / Fabricated or misquoted citations
Souza v. City of Fitchburg
U.S. District Court for the District of Massachusetts · Federal district courts · MA · · No. 23-cv-10140-DJC · other
On 2026-03-30, U.S. District Court for the District of Massachusetts dismissed the complaint. A self-represented party submitted an opposition brief with multiple fabricated or nonexistent authorities, including Mata v. Avianca, Inc., No. 22-cv-1461 and Allbaugh v. U. of Scranton, No. 3:24-CV-2237.
- Disposition
- The court dismissed the complaint and warned that generative AI may fabricate law, causes of action, and cases if a filer does not verify it.
- Recorded conduct
- The opposition brief used multiple fabricated or nonexistent authorities, including Mata v. Avianca, Inc., No. 22-cv-1461 and Allbaugh v. U. of Scranton, No. 3:24-CV-2237, which the court found nonexistent, misquoted, or unsupported. — outcome: warning; actor: litigant in person
- Canonical record
- Incident tracker entry — verified outcome, regulator disposition and related rules
What the court wrote
Verbatim from the court's document (public domain). Ellipses mark omitted text; nothing is paraphrased.
United States District Judge It is apparent to the Court that Plaintiff relied on generative AI in multiple instances to create his filings. [See e.g. ECF No. 8 at 4 (multiple instances where it appears the generative AI program is speaking directly to Plaintiff, using terms like “you” and “your” to refer to Plaintiff, and “he” to refer to defense counsel)]. While there is no rule requiring litigants to abstain from using generative AI, the Court warns Plaintiff that generative AI is unreliable for legal assistance and often “hallucinates,” where it either misstates the law or completely fabricates causes of actions and caselaw. See Mata v. Avianca, Inc., No. 22-cv-1461 (PKC), 2023 WL 3696209, at *2 (S.D.N.Y. May 4, 2023) (exemplifying generative AI-production of multiple fabricated case citations); see also Allbaugh v. U. of Scranton, No. 3:24-CV-2237, 2025 WL 2484188, at *6 (M.D. Pa. Aug. 28, 2025) (“Increased use of Artificial Intelligence (“AI”), particularly Generative AI (including, but not limited to, OpenAI's ChatGPT or Google's Bard), in the practice of law raises a number of practical concerns for the Court, including the risk that the generative AI tool might generate legally or factually incorrect information, or that it might create unsupported or nonexistent legal citations.”).