Special Report · The Law Skunk Standard
The Parade of
Horribles.
How AI-fabricated case law is ending legal careers across America — and why it will never happen on this platform. A report for citizens, lawyers, and law firms.
In June 2023, a federal judge in Manhattan sanctioned two lawyers for filing a brief containing six cases that never existed. An AI chatbot invented them — names, citations, quotations, holdings — and the lawyers filed them without checking. The case was Mata v. Avianca, and the profession told itself it was a fluke.
It was not a fluke. It was patient zero. A public database now tracks more than 120 court cases caught with AI-hallucinated authority. The documented record speaks for itself:
S.D.N.Y. — Mata v. Avianca (2023)
Six ChatGPT-invented cases filed in federal court.
$5,000 fine; lawyers ordered to notify their client and every judge falsely cited. The case that started it all.
Texas federal court (2024)
Nonexistent cases and AI-fabricated quotations.
$2,000 penalty plus mandatory generative-AI legal education.
California appellate matter (2025)
21 of 23 cited cases fabricated or carrying fictitious quotes.
$10,000 penalty.
Wyoming federal court (2025)
AI-drafted filing with non-existent cases in a high-profile defamation matter.
$3,000 per attorney.
High Court of Justice, London (2025)
18 of 45 citations fabricated in an £89 million case.
Court warned such conduct can constitute contempt — or criminal prosecution.
Kansas federal court (Feb. 2026)
Fictitious quotations and case references in a patent case.
$12,000 in combined fines.
Oregon federal court (May 2026)
Fabricated citations and quotations. The court surveyed the national record and called it a “parade of horribles.”
Approximately $109,700 in combined penalties and fees — among the largest to date.
Mississippi federal court (June 2026)
Four attorneys, one case, AI-generated fake citations.
Two attorneys barred for two years; all four removed from the case and fined.
9th U.S. Circuit Court of Appeals (June 2026)
Appellate briefs filled with fictitious cases.
$2,500 fines, six-month suspensions from appellate practice, forced AI-disclosure going forward.
11th U.S. Circuit Court of Appeals (July 2026)
“Fake and hallucinated” citations attributed to AI.
Public rebuke and warning against outsourcing legal work to machines unchecked.
Notice the trend line. Fines went from $5,000 in 2023 to six figures and multi-year bars by 2026. Courts now sanction lawyers for failing to catch opposing counsel's fake citations. Every sanctioned lawyer offered the same defense — "I didn't know the AI would lie to me" — and every court rejected it, because the duty of candor to the tribunal cannot be outsourced to a machine.
The AI never gets sanctioned. The human who didn't verify does.
To the Law Firms Reading This
If your associates use AI — and they do, whether you know it or not — your firm's name is one unchecked citation away from a published sanctions order. This platform was engineered so that cannot happen here. Every document our engines produce runs through an automatic citation audit: every case cite extracted, matched against a hand-verified library of over 200 landmark authorities, and anything unconfirmed sent through a second, independent review instructed to be ruthless \u2014 when in doubt, a citation is flagged, never blessed. You see the verdict on every cite before anything is filed: Verified, Verify Manually, or Fabrication Risk.
A lawyer who banks on this platform is never left fancy on his feet with no step. Every step is pre-calculated, every authority pre-verified, and the record stays clean — because in a courtroom the argument should win or lose on its merits. Never on its citations.
To the First Citizen
This platform was not ultimately built for law firms. It was built for a person we have not met yet: a citizen who is innocent until proven guilty — and who may be truly innocent — standing alone against the machinery of the state, without money, without counsel, without anyone who will listen. For that person, a single fabricated citation is not an embarrassment. It is the difference between credibility and erasure. Between going home and not going home. Between your children, your dog, your life in America — and losing all of it to a document that lied.
That is the level of integrity this platform holds itself to. Not because it is good marketing — because somebody's everything depends on it.
We will never tolerate, allow, or conduce a fabricated authority on this platform. Precisely dead on the money. Right. Correct. Period. No ifs, ands, or buts.
Sources: published sanctions orders and reporting by Reuters, The New York Times, The Guardian, Bloomberg Law, and the public AI-hallucination case database, 2023\u20132026. Case details reflect reporting available at publication; penalty figures vary in some accounts between trial and appellate proceedings.