By: Gabriella Varriale
As artificial intelligence (“AI”) continues to expand and develop, the legal field must continue to develop with it. While most lawyers and law firms have begun utilizing AI to some extent, there are still major risks involved with its usage. One of these risks is AI hallucinations in briefs that are prepared entirely by AI and not reviewed by the lawyer who is signing the brief, or any human being at all.
An AI hallucination occurs when an AI device creates a case citation that seems credible, but is false or does not exist entirely.[1] This could mean generating “. . . a citation to a case that does not contain the quoted language or generating a legal proposition that the cited authority never actually supported.”[2] These hallucinated citations are particularly dangerous in the litigation context because of the importance of case law in court.[3] Hallucinated case law may initially slip through the cracks due to how accurate the citations can be made to look.[4]
In July of 2026, the Court of Chancery was presented with a brief that contained AI hallucinations in Leiske v. Kidd.[5] This case involved a fee-advancement dispute that was originally before Magistrate Wright, who awarded advancement.[6] Vice Chancellor Will then reviewed the matter to decide exceptions to Magistrate Wright’s final report on advancement.[7]
On January 22, 2026, a prominent Delaware law firm that represented the plaintiffs filed an answering brief opposing the defendants’ exceptions.[8] Defendants then identified in their reply brief “what the Court described as ‘fictitious citations, fabricated quotations, and hallucinated legal propositions’ in the brief.”[9] The next day, on January 28, plaintiffs’ counsel filed a letter acknowledging the use of an AI tool to edit the test of their answering brief.[10] Plaintiff’s counsel “… admitted failing to verify the accuracy of the citations before filing, attributing the oversight to a paralegal’s review.”[11] Plaintiffs filed a corrected answering brief that removed quotation marks around erroneous statements of law but it did not correct the inaccurate descriptions of the courts’ analyses.[12] Defendant’s counsel then filed a letter detailing the remaining inaccuracies in the corrected answering brief and identifying similar citation problems in the plaintiffs’ summary judgment briefing.[13]
As a result, Vice Chancellor Will ordered the plaintiffs’ counsel and his firm to show cause as to why the Court should not issue sanctions for the faulty violations under Court of Chancery Rule 11(b) and 11(c).[14] Court of Chancery Rule 11(b) “…provides that an attorney or litigant presenting a pleading or written motion to the court certifies that ‘to the best of the person’s knowledge, information, and belief, formed after an inquiry reasonable under the circumstances[,]’ the ‘legal contentions’ made ‘are warranted by existing law.’”[15] Court of Chancery Rule 11(c) “mandates that ‘[a]bsent exceptional circumstances, a law firm must be held jointly responsible for violations committed by its partners, associates, or employees.’”[16] Vice Chancellor Will explained that “[b]ecause this incident may implicate the firm’s training, supervision, and deployment of GenAI, the firm must answer alongside the individual signatory.”[17]
The law firm and the lawyer who signed the briefs subsequently apologized to the Court for submitting the AI hallucinations by submitting affidavits to the Court, asking that sanctions not be imposed.[18] The law firm’s president “…said the firm has implemented a ‘GenAI-related program grounded in considered policy, informed by ethical responsibilities, and implemented through recurring discussion and training within the firm.’”[19] The lawyer who signed the briefs, who also helped lead the firm’s efforts to test AI tools, stated that he should have “‘…identified the necessary changes’ before the brief at issue was filed.”[20] It remains to be seen whether the plaintiffs will be sanctioned.
Although the plaintiffs in Leiske v. Kidd were ultimately successful in opposing the exceptions, their win was “… overshadowed by plaintiffs’ faulty reliance on GenAI.”[21] This serves as a cautionary tale for lawyers across the country: AI cannot be used without proper checks and balances. This does not mean that lawyers cannot use any type of AI tools to aid them but there must still be a thorough review by a human being before proceedings with the output. A lawyer or legal professional should review the output produced by the AI tool to ensure that the citations are accurate, the case law cited actually supports the contentions, and the overall product makes sense.
About the Author
Gabriella is a third-year regular division student at Widener University Delaware Law School and serves as the Web Editor for Volume 52 of the Delaware Journal of Corporate Law. During law school, Gabriella interned at the Delaware Court of Chancery for the Honorable Vice Chancellor Paul A. Fioravanti, Jr. and in the summer of 2026, she was a summer associate at Kelley Drye & Warren LLP in New York, New York. Gabriella is currently serving as a Josiah Oliver Wolcott fellow at the Delaware Supreme Court. Upon graduation, Gabriella will sit for the New York bar exam.
[1] Could My Own Law Firm’s AI Use Put My Case at Risk Without My Knowledge?, Primum Law Group, https://primumlaw.com/could-my-own-law-firms-ai-use-put-my-case-at-risk-without-my-knowledge/.
[2] Id.
[3] Id.
[4] Id.
[5] Letter Order, Leiske v. Kidd, 2025-0426-CDW (LWW) (Del. Ch. July 1, 2026).
[6] Francis G.X. Pileggi, Chancery Addresses AI Hallucinations in Court Filings, Del. Corp. & Com. Litig. Blog (July 23, 2026), https://www.delawarelitigation.com/2026/10/articles/chancery-court-updates/chancery-declines-to-award-remedy-for-failure-to-maintain-stock-ledger/.
[7] Id.
[8] Jaime A. Bartlett & Asher J. Zlotnik, How to Win and Still Lose: Delaware Court of Chancery’s Expanding Focus From Individual AI Misuse to Organizational Governance, Sidley: Enhanced Scrutiny (Sep. 24, 2026), https://ma-litigation.sidley.com/2026/09/how-to-win-and-still-lose-delaware-court-of-chancerys-expanding-focus-from-individual-ai-misuse-to-organizational-governance/.
[9] Id.
[10] Leiske, 2025-0426-CDW (LWW) (Del. Ch. July 1, 2026).
[11] Id.
[12] Id.
[13] Id.
[14] Leiske, 2025-0426-CDW (LWW) (Del. Ch. July 1, 2026).
[15] Id. (quoting Ct. Ch. R. 11(b)(2)).
[16] Id. (quoting Ct. Ch. R. 11(c)(1)).
[17] Leiske, 2025-0426-CDW (LWW) (quoting Ct. Ch. R. 11(c)(1)).
[18] Rose Krebs, Richards Layton Apologizes For AI Errors In Chancery Case, Law360 (July 16, 2026, 2:45 PM EDT), https://www.law360.com/articles/2501802/richards-layton-apologizes-for-ai-errors-in-chancery-case.
[19] Id.
[20] Id.
[21] Bartlett & Zlotnik, supra note 8.
Leave a Reply