United States (federal and state courts) · case law trend
AI Citation Sanctions: what it requires, and the records to keep
Since Mata v. Avianca in 2023, courts have sanctioned lawyers for filing briefs with fake or misdescribed authority produced by AI tools, under Rule 11, state equivalents and inherent power. Sanctions have escalated from small fines to public reprimands, disqualification, fee-shifting and bar referrals, so any AI-assisted drafting needs a human verification step before filing.
01 / What it requires
What AI Citation Sanctions requires
Lawyers and self-represented parties who sign and file court papers, and the firms that supervise them.
- 01Read and confirm every cited case, quotation and statute exists and supports the stated proposition before signing (Fed. R. Civ. P. 11(b)(2)).
- 02Treat AI output as unverified research; the signing lawyer stays personally responsible for every citation.
- 03Correct false citations promptly and candidly once discovered; do not compound them with false explanations to the court.
- 04Supervise associates, staff and contract lawyers who use AI tools, and enforce the firm's verification policy in practice.
- 05Follow judge-specific standing orders and local rules on disclosing or certifying AI use in filings.
02 / AI work
Where it touches AI work
The trend exists because generative AI tools can produce plausible but non-existent cases and quotations. Courts have held that using AI does not lower the signer's duty to verify, and a firm AI policy does not help if the verification step was skipped.
03 / Dates
The dates that matter
Mata v. Avianca (S.D.N.Y.): $5,000 Rule 11 sanction for ChatGPT-invented cases
Johnson v. Dunn (N.D. Ala.): three lawyers reprimanded, disqualified, referred to bar
Noland v. Land of the Free (Cal. Ct. App.) published opinion on AI-fabricated citations
Whiting v. City of Athens (6th Cir.): $15,000 fines per lawyer, reported
04 / Enforcement
Who enforces it
Courts, under Fed. R. Civ. P. 11, Fed. R. Bankr. P. 9011, 28 U.S.C. 1927, inherent authority and state rules; outcomes range from reprimand and fines ($5,000 in Mata; $15,000 per lawyer reported in Whiting, 6th Cir. 2026) to disqualification, fee awards and bar referral.
05 / Records
Records an organisation keeps
What an auditor or an inquiry usually asks to see. Your obligations depend on your role and your facts.
- 01Per-filing citation-check record showing who verified each authority
- 02Firm AI-use policy and dated training records
- 03List of approved AI tools and their vetting notes
- 04Supervision sign-off on briefs drafted with AI assistance
Optional evidence · where Planisphere fits
Keep a record anyone can check
When your team meets a duty, Planisphere can seal the event it reports as a receipt.
A disclosure shown, a review signed, a model change approved: the event your system reports.
The event becomes a sealed receipt, dated and attributed to whoever is signed in.
Paste the receipt into Verify. No account or key.
A record supports your compliance evidence process. It is not a legal finding or a certification. Systems that call Planisphere's gate API can also hold or refuse an action before it runs.
QUESTIONS
The questions that come next
Can I be sanctioned for AI-hallucinated citations?+
Yes. Courts regularly sanction lawyers who file briefs with fake or misdescribed authority produced by AI tools, using Rule 11, Bankruptcy Rule 9011, 28 U.S.C. 1927 and their inherent power. Penalties range from fines, such as $5,000 in Mata v. Avianca (2023) and $15,000 per lawyer in Whiting v. City of Athens (6th Cir. 2026), to public reprimand, disqualification from the case and referral to the bar. Using an AI tool is no defense: the lawyer who signs a filing must confirm every citation in it. Read the full answer
What was Johnson v. Dunn and what does it mean for AI legal tools?+
Johnson v. Dunn is a July 23, 2025 sanctions order from the federal court in the Northern District of Alabama. Three Butler Snow lawyers defending a former Alabama prisons commissioner filed motions citing cases that ChatGPT had invented. The court said reprimands and modest fines were not deterring this conduct, so it publicly reprimanded the three lawyers, removed them from the case, ordered the opinion published and sent it to the Alabama State Bar. The lesson for AI legal tools is that the signing lawyer, not the tool, answers for every citation. Read the full answer
Does having an AI policy protect my firm from sanctions?+
No. A written AI policy helps show the firm supervises its lawyers, but it does not stop a court from sanctioning lawyers who file unverified citations. In Johnson v. Dunn (N.D. Ala. 2025), Butler Snow's policy required written approval before anyone used ChatGPT for legal research. One lawyer ignored it, and the court sanctioned him and two colleagues; the firm itself was released without sanction. What protects clients and the firm is a policy that is followed in practice: training, supervision under Model Rules 5.1 and 5.3, and a verification step before every filing. Read the full answer
How do I verify AI legal research before filing?+
Check every authority yourself in a reliable legal research service before you sign. Confirm that each case or statute exists, that the cited page and any quotation appear in it, that it supports the point you cite it for, and that it is still good law. Do not rely on the AI tool's own summary or on a second AI tool to check the first. Keep a note of who checked each citation and when. Courts treat verification as the signing lawyer's personal duty, and supervising lawyers must make sure staff follow it. Read the full answer
SOURCES
Go back to the source
U.S. District Court, S.D.N.Y. (RECAP copy via CourtListener)
U.S. District Court, N.D. Ala. (docket copy via Docket Alarm)
Administrative Office of the U.S. Courts
CHANGE LOG
What changed, and when
Rebuilt from primary sources; facts checked on this date.
Related: AI Citation Sanctions for law AI workflows · SBOM · C-SCRM · Proctoring Fairness