Vice Chancellor Lori W. Will found that counsel's corrected brief merely deleted the quotation marks while preserving the fabricated legal synthesis. Now Richards, Layton & Finger, the firm and not just the partner who signed, has until July 15 to put its written GenAI policies on a public docket under Rule 11's joint-responsibility clause.
On July 1, 2026, Vice Chancellor Lori W. Will of the Delaware Court of Chancery put a law firm's correction on the record for what it was. The finding came in her letter ruling in Leiske v. Kidd, C.A. No. 2025-0426-CDW, an advancement case in Wilmington. "In correcting the Answering Brief, counsel merely deleted quotation marks while preserving the fabricated legal synthesis." Now Richard P. Rollo, the partner who signed that brief, and Richards, Layton & Finger, the firm behind the signature, have until July 15, 2026 to show cause why they should not be sanctioned. Not just the lawyer. The whole firm.
Sanctions fights over AI-invented citations are no longer rare. A tracker maintained by researcher Damien Charlotin counts 1,693 cases of hallucinated authority so far, in courts from Delaware to the Supreme Court of India. Most end the same way. A fine for the lawyer who filed, an apology on the docket, a next case. Vice Chancellor Will's ruling moves the target. She ordered the firm itself to answer alongside its partner, and she ordered the firm's own AI governance into the answer.
The mechanism is Court of Chancery Rule 11(c)(1), quoted in the ruling: "[a]bsent exceptional circumstances, a law firm must be held jointly responsible for violations committed by its partners, associates, or employees." By July 15, an authorized representative of Richards Layton must file an affidavit describing the firm's written GenAI policies in effect during December 2025 and January 2026, how those policies were communicated to Rollo and the team on this matter, the training programs and safeguards the firm has implemented or plans to implement, and, in the ruling's words at page 8, "[a] showing of 'exceptional circumstances,' if any exist, demonstrating why the law firm should not be held jointly responsible for any Rule 11 violations committed by its partner." When that affidavit lands, a firm's internal AI policies will sit on a public docket, readable by any client, any adversary, and anyone with a CourtListener account.
The path to that order runs through five days in January. On January 22, plaintiffs' counsel filed the Answering Brief. On January 27, the defense reply flagged fictitious citations, fabricated quotations, and hallucinated legal propositions inside it. On January 28, counsel filed a letter acknowledging that a GenAI tool had been used to revise the brief, attributed the missed verification to a paralegal's review, and filed a corrected version. That version removed the quotation marks around erroneous statements of law and left the descriptions standing. The corrected brief still cited Perconti v. Thornton Oil Corp. for a principle the Vice Chancellor found nowhere in Perconti. On January 29, counsel wrote again, this time faulting the defense for failing to meet and confer before alerting the court, insisting the cited authorities "are real" and "support the arguments advanced," and accusing the other side of using the errors for a "litigation advantage."
Trial drama, nuclear verdicts, and the plaintiff-firm tactics behind them. Court-reporter prose, no consultant filler. Read by litigation leaders at F500 legal departments and national carriers. Free.
The July 1 ruling answers each move in turn. On the paralegal: "Counsel cannot discharge those obligations by relying on a paralegal, much less on an artificial intelligence program." On the demand to meet and confer first: "There is nothing to negotiate or compromise when an attorney submits false citations to a tribunal." Rollo's own affidavit, due the same July 15, must reconstruct who entered the prompts, what cite-checking happened before filing, whether any attorney verified the quoted text against the actual opinions, and why the quotation marks came out while the flawed legal propositions stayed in.
The ruling places itself in a hardening line of authority. It cites Mata v. Avianca, the 2023 New York case holding that "[a] fake opinion is not 'existing law,'" and Johnson v. Dunn, where an Alabama federal court disqualified the lawyer outright. Judge Mark D. Clarke, last seen in these pages sanctioning an Oregon filing he called "a notorious outlier in both degree and volume" at a cost of $110,000, aimed his order at the lawyers who filed the fabrications. Vice Chancellor Will aimed one level up, at supervision, training, and the written policy on the shelf.
The next fight sits on a clock. The advancement merits are already decided; she overruled the defendants' exceptions on February 2 and expressly retained jurisdiction "to determine what remedy, if any, should issue" for the GenAI errors. That reservation is why this collateral fight outlived the case that produced it. Once the affidavits arrive, she will decide whether the Rule to Show Cause resolves on the papers or moves to an evidentiary hearing or oral argument in Wilmington.
The correction took one day. The explanation gets two weeks, an oath, and the firm's name at the top. On July 15 the affidavits come due on the docket, and the Vice Chancellor reads what Richards Layton wrote down about AI before this case and what it plans to write now. The quotation marks are gone. The questions are not.
The Executive Briefing is six questions. It shows you exactly where the gaps are.
Take the Executive Briefing →Trial drama, nuclear verdicts, and the plaintiff-firm tactics behind them. Court-reporter prose, no consultant filler. Read by litigation leaders at F500 legal departments and national carriers. Free.