In February 2026, Omaha attorney Greg Lake filed a brief in a divorce appeal. The brief contained 57 of 63 defective citations, including 20 hallucinated cases, three fabricated decisions, and invented statutory quotes. When a judge asked him directly whether he had used AI to write it, Lake first denied it, then admitted he had and called it a grave error of judgment. Two months later, the Nebraska Supreme Court suspended him until further notice and ordered a disciplinary investigation. This is being reported as one of the first U.S. bar disciplines to remove a lawyer from practice entirely over AI filings.
Lake isn't an outlier and he isn't the last. A public database maintained by legal researcher Damien Charlotin now tracks roughly 1,490 court decisions worldwide, more than 1,000 of them in the United States as of May 2026, where a party relied on AI-hallucinated material and a court responded. This didn't start in a vacuum. The original cautionary tale dates back to 2023, when a New York lawyer submitted six fabricated ChatGPT-generated citations against an airline and was fined $5,000. Three years later, that fine looks quaint: penalties have climbed from four-figure fines to $15,000 per attorney in a federal court of appeals, plus the first bar suspensions tied to AI filings.
What makes these cases worse isn't usually the original mistake — it's what happens after. What drove the severity in Lake's case was the denial; candor with the court is the one thing an AI cannot do for the lawyer who used it. Across the tracked cases, three patterns repeat, and the cover-up consistently draws the harsher penalty. This isn't unique to solo practitioners either. A June 2026 Mississippi federal case, Withers v. City of Aberdeen, saw all four lawyers involved removed from the case after nobody took responsibility for AI-hallucinated material in their filings. The technology failure is the cheap part. The institutional failure to catch it before it reaches a judge is what ends careers.
None of this means legal AI tools are useless — most large firms and solo practices alike are using them for research, drafting, and summarization every day. The failure mode is specific and fixable: nobody built a verification step between the AI's output and the filed document. A citation-checking pass takes minutes. Skipping it took a Nebraska lawyer's license.
If your firm or your company's legal team uses AI for research, drafting, or contract review, the fix isn't banning the tool — it's assigning a named human to verify every citation, case reference, and factual claim before anything goes out the door, and documenting that they did it. If you can't say who checked the last AI-assisted filing your firm sent to a court, that's the gap to close this week, not next quarter.