By Jane Susskind
Artificial intelligence (“AI”) promises to transform how we practice law. But while AI-powered tools can increase efficiency, they also introduce risks and ethical concerns. Recognizing these risks, the ABA Standing Committee on Ethics and Professional Responsibility issued its first opinion in July 2024 on the use of generative AI and many judges across the country have established protocols for using AI in their courtrooms. For example, Judge Martínez-Olguín of the Northern District of California instructs in her standing order that “[a]ny submission containing AI-generated content must include a certification that lead trial counsel has personally verified the content’s accuracy.” Standing Order for Civil Cases Before District Judge Araceli Martínez-Olguín, https://www.cand.uscourts.gov/wp-content/uploads/2023/03/AMO-Civil-Standing-Order-11.22.2023-FINAL.pdf (last accessed Apr. 29, 2025). Law firms are following suit, regularly updating their AI policies to keep up with AI’s adoption in the legal profession.
Even with this guidance, attorneys have landed in the national spotlight—not for persuasive legal writing or skilled advocacy, but for submitting briefs with AI-generated errors. By now, we’ve all seen news headlines and heard of the phrase “AI hallucinations,” which refers to false information (e.g., non-existent cases, holdings, and quotations) generated by large language model systems. Reviewing a few of these cases can help us understand the scope of sanctions for improperly relying on AI, as well as remedial steps we can take to avoid sanctions. Most importantly, these cases provide guidance for attorneys who want to efficiently and ethically incorporate AI into their practice.
Scope of Sanctions
It’s important to consider the scope of sanctions for submitting briefs with AI-generated hallucinations and errors. In one of the first decisions addressing these concerns, a New York federal judge imposed a $5,000 fine on attorneys who submitted a brief with AI-generated fake cases. Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023). Since then, courts have not shied away from imposing monetary sanctions for similar misconduct. See, e.g., Gauthier v. Goodyear Tire & Rubber Co., No. 1:23-CV-281, 2024 WL 4882651 (E.D. Tex. 2024) (imposing a $2,000 sanction); Wadsworth v. Walmart Inc., 348 F.R.D. 489 (D. Wyo. 2025) (imposing a $3,000 sanction); Mid Cent. Operating Eng’rs Health & Welfare Fund v. HoosierVac LLC, No. 2:24-CV-00326-JPH-MJD, 2025 WL 574234 (S.D. Ind. 2025) (recommending a $15,000 sanction).
But monetary sanctions aren’t the only sanctions attorneys face. In February of this year, a Wyoming federal judge revoked an attorney’s pro hac vice admission after the attorney filed a motion in which eight of nine cases cited were AI hallucinations. Wadsworth, 348 F.R.D. 489. Some judges have ordered attorneys to provide written notice of their misconduct and apologize to their clients and the judges wrongly attributed to the AI-generated cases. Mata, 678 F. Supp. 3d 443. Other judges turn to social media to deter the improper use of AI. For example, Texas Judge Roy Ferguson posted an appellate court’s order on X, warning: “Always check your cites! This is why I have a standing order on the use of AI. The underlying appellate brief contains fabricated and nonexistent citations.” Lauren Berg, “Texas Appeals Court Calls Out Seemingly AI-Generated Cites,” Law360, July 27, 2023. And at least one judge has suggested that professional discipline might be the appropriate sanction. See Mid Cent., 2025 WL 574234.
Remedial Steps
Now that we know what’s at stake, what can we do to avoid sanctions? First and foremost, always check and verify the sources generated by AI for accuracy. Attorneys have a professional duty and ethical obligation to ensure that all legal authority they present to the court is good law. As one judge aptly summarized, “[w]hile technology continues to evolve, one thing remains the same—checking and verifying the source.” Wadsworth, 348 F.R.D. 493. We should also be aware of AI’s shortcomings and limitations. The most well-known shortcoming is AI’s capacity to “hallucinate” or make up legal authority. But there are others, like AI’s inability to apply true legal reasoning and judgment. If you plan on adopting AI into your practice, consider reading up on these shortcomings and attending courses and CLEs on the use of generative AI in the legal profession.
And if you find yourself in the unfortunate position of having unknowingly submitted a motion with AI-generated mistakes, try to remedy it as soon as possible. Promptly withdraw the motion or file a supplement providing correct citations and legal authorities. Be honest about your use of AI and apologize to the court and opposing counsel. You may even consider offering to reimburse the opposing party for fees and costs incurred in uncovering the improper use of AI. These remedial efforts can go a long way with the court. See, e.g., Wadsworth, 348 F.R.D. 489 (appreciating counsel’s “remedial steps, transparency, and apologetic sentiments” and recommending that attorneys follow these steps to avoid sanctions); Mata, 678 F. Supp. 3d 443 (suggesting that had the attorneys “com[e] clean” sooner, they would not have been sanctioned).
The Takeaway?
AI has the potential to transform how we research and write briefs, but it has a long way to go. AI may be a starting point for legal research, but it should not be the endpoint. Always check and verify your sources and stay up to date on AI’s advancements and limitations. As Magistrate Judge Dinsmore advised, “the use of artificial intelligence must be accompanied by the application of actual intelligence in its execution.” Mid Cent., 2025 WL 574234 (emphasis added). This is good advice we can all take to heart.
Disclosure: The author used Perplexity, an AI tool, to find Judge Martínez-Olguín’s standing order on the use of AI-generated content in her courtroom. The author verified the accuracy of the information by visiting the United States District Court for the Northern District of California’ website and independently locating the standing order. The author did not use AI to analyze the order or write any portion of this article.
About the author
Jane Susskind is an Associate at McDonald Carano. She is a member of the Commercial & Complex Litigation Practice Group and Appellate Practice Group. Before joining McDonald Carano, Jane served as a judicial law clerk for Nevada Supreme Court Justices Ron Parraguirre, Elissa Cadish, and Mark Gibbons.
About the article
This article was originally published in the Communiqué (June/July 2025), the official publication of the Clark County Bar Association. See https://clarkcountybar.org/about/member-benefits/communique-2025/communique-june-july-2025/. The printed magazine was mailed to CCBA members on June 6, 2025.
The articles and advertisements appearing in Communiqué magazine do not necessarily reflect the opinion of the CCBA, the CCBA Publications Committee, the editorial board, or the other authors. All legal and other issues discussed are not for the purpose of answering specific legal questions. Attorneys and others are strongly advised to independently research all issues.
© 2025 Clark County Bar Association (CCBA). All rights reserved. No reproduction of any portion of this issue is allowed without written permission from the publisher. Editorial policy available upon request.