What real attorneys are saying about AI-hallucinated citations

17+ verbatim quotes from public sources. Each quote links to the original — a court opinion, sworn affidavit, published bar-association guidance, or on-the-record legal press coverage. Not our copy. If a quote is even slightly paraphrased from the underlying source it is marked below.

• “Greatly regrets having utilized generative artificial intelligence to supplement the legal research performed herein and will never do so in the future without absolute verification of its authenticity.”

  Steven A. Schwartz, Esq. (Levidow, Levidow & Oberman, P.A.) — affidavit filed in Mata v. Avianca, as quoted by CNN

  CNN Business · 2023-05-27paraphrased

• “The Court is presented with an unprecedented circumstance. A submission filed by plaintiff's counsel in opposition to a motion to dismiss is replete with citations to non-existent cases.”

  Hon. P. Kevin Castel, U.S. District Judge (S.D.N.Y.) — Order to Show Cause, Mata v. Avianca

  CourtListener — Mata v. Avianca docket · 2023-05-04

• “Existing rules impose a gatekeeping role on attorneys to ensure the accuracy of their filings. [Counsel] abandoned their responsibilities when they submitted non-existent judicial opinions with fake quotes and citations created by the artificial intelligence tool ChatGPT.”

  Hon. P. Kevin Castel — sanctions opinion, Mata v. Avianca, 678 F. Supp. 3d 443 (S.D.N.Y. 2023)

  CourtListener — Mata v. Avianca opinion · 2023-06-22

• “At the very least, the duties imposed by Federal Rule of Appellate Procedure 46(c) require that attorneys read, and thereby confirm the existence and validity of, the legal authorities on which they rely.”

  Second Circuit panel (per curiam) — Park v. Kim, 91 F.4th 610 (2d Cir. 2024)

  FindLaw — Park v. Kim (2d Cir. 2024) · 2024-01-30

• “Filing an appellate brief with bogus citations in this Court for any reason cannot be countenanced and represents a flagrant violation of the duties of candor Appellant owes to this Court.”

  Missouri Court of Appeals, Eastern District — Kruse v. Karlen, 692 S.W.3d 43 (2024)

  FindLaw — Kruse v. Karlen (Mo. App. E.D. 2024) · 2024-02-13

• “As attorneys transition to the world of AI, the duty to check their sources and make a reasonable inquiry into existing law remains unchanged.”

  Chief Magistrate Judge Kelly H. Rankin — Wadsworth v. Walmart, Inc. (D. Wyo. Feb. 24, 2025)

  CourtListener — Wadsworth v. Walmart sanctions order · 2025-02-24

• “Fines do not account for the extreme dereliction of professional responsibility that fabricating citations reflects, nor for the many harms it causes.”

  U.S. District Court, N.D. Alabama — Johnson v. Dunn sanctions order (2025)

  Berkeley Law — Johnson v. Dunn, 792 F. Supp. 3d 1241 (N.D. Ala. 2025) · 2025-07-23

• “In the quickly expanding universe of cases involving sanctions for the misuse of artificial intelligence, this case is a notorious outlier in both degree and volume.”

  U.S. Magistrate Judge Mark D. Clarke (D. Or.) — $110,000 sanction order, as reported by ABA Journal

  ABA Journal · 2026-04-17

• “Another ChatGPT lawyer, Richard Bednar, was sanctioned by the Utah court of appeals after filing a legal brief with false citations.”

  Thomson Reuters Legal blog — coverage of Utah Bednar sanction

  Thomson Reuters Legal · 2025-08-06paraphrased

• “Among respondents to 2025 Generative AI in Professional Services Report who said they felt GenAI should not be a part of their daily work, 40% cited accuracy and reliability as their primary concern — nearly double any other major concern, including a lack of human touch (22%), generality of outputs (19%), or biased data (12%).”

  Thomson Reuters Institute — GenAI hallucinations coverage (Zach Warren)

  Thomson Reuters Institute · 2025-08-18

• “Considering either free or paid versions of ChatGPT in the first place should raise major flags among small law firms trying to get ahead while balancing trust and innovation. A professional-grade legal AI solution should cite cases and pull clauses that can hold up to legal scrutiny.”

  Thomson Reuters Legal blog — small-firm ChatGPT guidance (Abby Colson)

  Thomson Reuters Legal · 2025-08-06

• “Our previous study of general-purpose chatbots found that they hallucinated between 58% and 82% of the time on legal queries, highlighting the risks of incorporating AI into legal practice. […] the Lexis+ AI and Ask Practical Law AI systems produced incorrect information more than 17% of the time, while Westlaw's AI-Assisted Research hallucinated more than 34% of the time.”

  Stanford RegLab and Stanford HAI — Magesh, Surani, Dahl et al.

  Stanford HAI · 2024-05-23

• “ABA Formal Opinion 512 is the ABA's first formal ethics guidance on generative AI tools. It reiterates that lawyers using GAI must fully consider their applicable ethical obligations, including duties of competence, confidentiality, communication with clients, and maintaining candor toward tribunals.”

  American Bar Association Standing Committee on Ethics and Professional Responsibility (news release)

  American Bar Association · 2024-07-29paraphrased

• “Consumer-facing AI tools like ChatGPT operate on open networks that create unacceptable risks for legal professionals. This exposure violates fundamental professional responsibilities for attorneys bound by strict confidentiality rules and ethical obligations.”

  Thomson Reuters Legal blog — Why small law firms need Fiduciary-Grade AI (Sabrina Corsiga)

  Thomson Reuters Legal · 2025-08-07

• “OpenAI, the company behind ChatGPT, is now required to keep user chat logs — even deleted ones — because of a court order in the New York Times v. OpenAI case. This means that for anyone using Free, Plus/Pro, or Team plans, your chats could be stored indefinitely, even if you delete them or use the "temporary chat" feature.”

  Bar Association of San Francisco — attorney-facing privacy notice (Amy Jo McGuigan)

  Bar Association of San Francisco · 2025-08-07

• “The AI Hallucination Cases database — maintained by Damien Charlotin, a researcher at HEC Paris, a leading business school in France — identifies a half-dozen federal court filings in Arizona since September 2024 that include fabricated material from ChatGPT or another generative AI tool. […] Worldwide, the database shows 486 cases — 324 of them in U.S. federal, state and tribal courts.”

  Cronkite News — Tufan Neupane, summarizing Damien Charlotin's public database

  Cronkite News (Arizona PBS) · 2025-10-28

• “Confronted with AI hallucinations in filings, one court shows 'justifiable kindness,' while another gets tough.”

  ABA Journal — Debra Cassens Weiss, coverage of Hall v. Academy Charter School (E.D.N.Y.) and Mavy v. Commissioner of Social Security Administration (D. Ariz.)

  ABA Journal · 2025-08-19