AI-hallucinated citations delay wage class action settlement in N.D. Cal

Tombstone icon

A federal judge in the Northern District of California sanctioned plaintiff's counsel James Dal Bon in Buchanan v. Vuori Inc. (Case 5:23-cv-01121-NC) for filing AI-generated case law citations in a motion for preliminary approval of a wage and hour class action settlement. Dal Bon used six different AI tools to prepare the memorandum, which contained hallucinated quotes and a nonexistent case citation. After the court flagged the fabricated citations, his corrected filing still contained AI-hallucinated case law. The sanctions delayed the class action settlement, ultimately converting it to an individual settlement that abandoned the class members the attorney was supposed to represent.

Incident Details

Severity:Facepalm
Company:ChatGPT users (law firm)
Perpetrator:AI chatbot
Incident Date:
Blast Radius:Class action plaintiffs whose settlement was delayed; attorney sanctioned for AI-generated fabrications that persisted even after correction

The Case

Terrence Buchanan filed a wage and hour class action lawsuit against Vuori Inc., a clothing company, alleging the company underpaid him and other class members for overtime work and bonuses earned during his employment. The case, 5:23-cv-01121-NC, was filed in the U.S. District Court for the Northern District of California before Magistrate Judge Nathanael M. Cousins.

The case itself was unremarkable - wage and hour violations are among the most common class actions in federal court. The parties reached a settlement agreement, and plaintiff's counsel needed to file a motion for preliminary approval of the class settlement. This is a procedural step where the judge reviews the proposed settlement to make sure it's fair to the class members before giving the go-ahead. It's usually straightforward.

James Dal Bon, a solo practitioner at Wisdom Law Group APC, was class counsel for the plaintiff. He submitted his memorandum of law in support of preliminary approval. That's when things fell apart.

Six AI Tools

"As a solo practitioner under time pressure, Dal Bon admits that he used about six different AI tools to prepare his motion," Judge Cousins noted in his sanctions order.

Six AI tools on a single brief. The resulting memorandum was - in the judge's words - "rife with hallucinated quotes and one nonexistent case citation." The AI-generated content didn't just include one bad citation. The document contained fabricated quotations attributed to real cases, quotes that those cases never actually said. And at least one case citation pointed to a case that didn't exist at all.

On November 5, 2025, after discovering the fabrications, Judge Cousins ordered Dal Bon to show cause as to why he should not be sanctioned under Federal Rule of Civil Procedure 11(c) and referred to the court's Standing Committee on Professional Conduct under Civil Local Rule 11-6.

The Failed Correction

Getting caught using AI to fabricate citations is bad. What happened next was worse.

Dal Bon submitted a corrected memorandum with a second motion for preliminary approval. The corrected version still contained AI-hallucinated case law.

The attorney had been put on notice that his AI tools were generating fabricated legal authorities. He was facing a show cause order for that exact problem. And the corrected filing he submitted to fix the problem still had the same category of error. The court had given him a chance to clean up his work, and the cleanup was also defective.

Judge Cousins cited this in his November 20, 2025 sanctions order specifically: the corrected memorandum "did not correct the false case law hallucinated by the AI tools." The repeat failure made a decisive difference in how the court handled the case.

The Sanctions

Dal Bon offered the court several proposed remedies: he would forfeit his attorneys' fees in the matter, file an amended motion certifying that he had verified all citations, and complete continuing legal education courses. Judge Cousins declined all of these proposals.

The court imposed its own sanctions:

  1. Plaintiffs' second motion for preliminary approval of the class action settlement and the corrected motion were stricken with prejudice - meaning they could not be refiled.
  2. Dal Bon was ordered to pay $250 to the clerk of court by December 5, 2025, under Rule 11(c).
  3. Dal Bon was referred to the court's Standing Committee on Professional Conduct in connection with his violation of Local Rule 11-4 and unprofessional conduct.

The $250 fine was the smallest part of the order. The striking of the motions with prejudice and the referral to professional conduct proceedings were far more consequential.

Judge Cousins referenced a growing body of case law on AI sanctions: "Courts that have sanctioned similar violations of Rule 11(b) - namely, filing hallucinated or otherwise false case citations and quotations - have imposed a range of sanctions, often in combination, including monetary sanctions, striking the filing containing erroneous citations, requiring written notification to the client, requiring written notification to judges incorrectly identified as having authored nonexistent cases, and referring attorneys to the appropriate disciplinary body for disciplinary proceedings."

Inadequate Class Counsel

The most significant finding was Judge Cousins' determination that he "cannot find Dal Bon is adequate class counsel based on his representation in the case thus far." This finding, while stopping short of formally removing Dal Bon from the case, effectively killed the class action.

Class actions require adequate representation. The named plaintiff represents thousands of absent class members who are relying on their attorney to handle the case competently. If the court finds that class counsel is inadequate, the class can't be certified, and any class settlement can't be approved. Without class certification, there is no class action.

The sanctions prohibited Dal Bon from filing any further motions to approve a class settlement. The class members - Vuori workers allegedly owed unpaid wages and bonuses - couldn't get their settlement approved because their own attorney's AI-drafted brief torpedoed the proceeding.

The Individual Settlement

On January 7, 2026, counsel for Vuori filed a case management statement explaining the aftermath. The parties had signed a settlement agreement containing "a general release of Plaintiff's claims against Defendant," and per the terms of that agreement, Buchanan was obligated to dismiss the action with prejudice by December 31, 2025.

On January 9, 2026, Judge Cousins granted a joint dismissal. The case ended with a confidential individual settlement between Buchanan and Vuori. Dal Bon signed the dismissal motion alongside Vuori's counsel.

The critical difference: the class action settlement would have covered all the workers affected by Vuori's alleged wage practices. The individual settlement covered only Terrence Buchanan. The other class members - however many workers had similar claims against Vuori - got nothing. Their claims were not released by Buchanan's individual settlement, but they also no longer had a pending class action to join or benefit from.

The Solo Practitioner Problem

Dal Bon's case illustrates a specific risk pattern in AI-assisted legal work. Solo practitioners and small firms face the same filing deadlines and procedural requirements as large firms, with a fraction of the resources. AI tools promise to close that gap - generate a brief in hours instead of days, produce research that would take associates at a large firm to compile.

The problem is that the verification step doesn't scale down. A large firm has associates to cross-check citations, partners to review drafts, and librarians to verify sources. A solo practitioner using six AI tools to generate a brief is also the only person who can verify the output. When the time pressure that led to using AI in the first place also prevents thorough verification, the attorney is submitting AI-generated content on trust.

Dal Bon said he was "under time pressure." The AI tools were supposed to help him meet that pressure. Instead, they introduced fabrications that he did not catch, and when given a second chance, he still did not catch them. The tools didn't solve his capacity problem. They transformed it from "not enough time to write a brief" into "not enough time to verify a brief I didn't write."

Impact on the Plaintiffs

The Duane Morris analysis of Buchanan highlighted an implication for class action defense: employers should "review class counsel's filings scrupulously by noting any citations or quotations that seem incorrect and AI-generated as this may build a case for disqualifying class counsel and may prove as a barrier to getting approval of a class settlement agreement."

That's the defense bar reading this decision as a new tactical tool. If class counsel uses AI and the output contains errors, the defense can move to disqualify that counsel and potentially kill the class action. In effect, one attorney's AI error doesn't just sanction that attorney - it can destroy the case for the entire class.

The workers in Buchanan v. Vuori alleged they were underpaid. Their attorney used AI to save time. The AI generated fabricated law. The attorney missed it twice. The court struck the settlement and found counsel inadequate. The class action collapsed into an individual settlement. The workers' claims survived in theory but lost their vehicle in practice.

Discussion