The first quarter of 2026 marks the moment US courts stopped issuing warnings and started writing checks. Across federal and state courts, judges imposed more than $145,000 in sanctions against attorneys who submitted briefs containing AI-fabricated case citations, according to a Q1 2026 enforcement review. The single largest penalty so far was $96,000, levied against a New York attorney earlier this month.
That attorney is Stephen Brigandi, sanctioned by US Magistrate Judge Mark Clarke on April 4 after his court filings contained 23 fabricated legal citations and 8 false quotations across three submissions. The total court-imposed penalties in his case exceeded $110,000 once procedural costs were added. Reason's Volokh Conspiracy noted that this is now the second-largest individual sanction for AI-related court misconduct on record, behind only a separate $145,000 ruling tied to a class-action filing earlier this year.
From warnings to bills
The escalation pattern is striking. Total court sanctions for AI-fabricated citations were $5,000 in January, $250 in February, then jumped to over $139,000 in March alone. By the end of April, courts in the Sixth Circuit, Oregon, and several state systems had all issued penalties in the same window. Researcher Damien Charlotin, who maintains a global database of AI hallucination cases in legal filings, has now catalogued 1,227 cases worldwide.
The Sixth Circuit issued one of the more consequential rulings of the quarter, sanctioning two attorneys $30,000 in combined penalties for more than two dozen fake citations and dismissing their case entirely for what the court called "pervasive misconduct." The Sixth Circuit Appellate Blog noted that the ruling matters less for the dollar figure than for the precedent: dismissing the case removed the attorneys' clients from court entirely.
The Nebraska suspension
The most severe consequence of the quarter was professional. The Nebraska Supreme Court suspended attorney Greg Lake from practicing law after his appellate brief contained 57 defective citations. Suspension is a meaningfully different sanction from a fine. It removes the attorney's ability to earn a living from law for the duration of the suspension, and it must be disclosed to potential employers and clients afterward.
The judges' position is consistent
Courts have converged on a clear standard: attorneys have a non-delegable duty to verify every citation in any document they sign, regardless of whether the source was a junior associate, a paralegal, or an AI tool. The reason is that filing fabricated citations wastes opposing counsel's time, wastes judicial time, and corrupts the case-law record itself. Courts cite future cases by reference, and a fabricated citation slipped into one ruling can propagate forward.
What makes the AI version of this problem distinctive is the volume. Where a junior associate might fabricate one citation, a model can produce twenty in a single response without distinguishing real cases from invented ones. Researcher Charlotin's database shows the median sanction has grown roughly tenfold since he started tracking it in late 2024, and the share of cases with fabricated citations counted in the dozens rather than the handful is rising.
The $145,000 paradox
One thread running through this quarter's coverage is what The Ethics Reporter called the $145,000 paradox: 61% of federal judges in a recent survey reported using AI tools themselves for research and drafting, while sanctioning attorneys for the same general category of mistake. The judicial position is consistent on its own terms, since judges verify their own work, but the optics of the gap are uncomfortable, and several bar associations are now revisiting their AI use guidance.
The practical outcome of Q1's sanctions is that AI use in legal filings is no longer a question of policy debate. It is a question of malpractice insurance. Several major firms have moved to mandatory citation-verification protocols that include screenshots of every cited case, and at least one insurance carrier has begun requiring AI-use disclosures on professional liability renewals.
Commentarii · 0