FEATURE ARTICLE: Ethical Issues with Judicial Use of Generative AI
Author: Professor Michael H. Hoeflich, PhD, Editor-in-Chief
Legal Editor: Carrie E. Parker
This article is featured in Volume 7, Number 8 of the Legal Ethics and Malpractice Reporter, published August 31, 2026.
Lawyers have been using Generative Artificial Intelligence (GAI) for the past decade. During that time, AI has evolved rapidly, and dependence on AI by lawyers has increased enormously. With that increased usage, lawyers have discovered that there are often critical problems with GAI.
When used to do research and draft documents such as memoranda, motions, and briefs, GAI makes mistakes and cites fictional cases and other sources, called hallucinations, and misinterprets real cases that it cites. Indeed, the number and quality of lawyers and law firms that have submitted flawed GAI-generated documents to courts is rather frightening.
Damien Charlotin maintains a website that keeps track of hallucination cases.1 According to his records, there have been close to 2,000 cases in which GAI involvement has been discovered in legal documents as a result of hallucinations being submitted for official use. Lawyers who have submitted these documents have ranged from solo practitioners to members of the most elite law firms. Even though the problem has become well known throughout the Bar, this practice is not decreasing. Meanwhile, the penalties imposed on the lawyers have been significantly increasing.
Virtually all the attention about legal use of GAI has focused on hallucinations in documents submitted to courts or produced for courts and how procedural rules and ethical rules may be violated by the lawyers. But this is only one part of the problem.
Now the profession is beginning to consider the problem of how to deal with hallucinations which make it into judicial opinions and orders—either because a judge or judicial clerk has used GAI or because the judge or judicial clerk has repeated hallucinations submitted by an attorney.
It is important to realize that judges—lots of them—are using GAI. Sen. Charles Grassley held a Senate hearing on this issue in October 2025, and, in a speech on the Senate floor, expressed his concern that federal judicial use of GAI was a serious issue for the federal government:
This month, I sent letters to a pair of federal judges regarding their suspected use of generative Artificial Intelligence (AI) in published court orders.
The two judges issued orders that were wrought with serious, inexplicable errors that were of great concern to the litigants, to the public and to me.
One judge, sitting in the Southern District of Mississippi, issued an order that named plaintiffs and defendants who were not even parties to the case.
It misquoted statutory text. It erroneously cited evidence that didn’t exist in the record. And it referred to sworn declarations by people who had not appeared anywhere in the case.
I’ve never seen or heard of anything like this from a federal court.
The other judge, sitting in the District of New Jersey, issued an order with fake quotes and misstatements about case law.
The order even falsely attributed quotations to the defendants that apparently were never made.
The two cases Grassley mentioned were obvious egregious examples of the types of errors that can be caused by unverified use of GAI, but they are not the only cases where errors attributable to GAI have been entered into the court record by the judge. Charlotin’s webpage shows that other U.S. judges have had similar problems with unverified GAI and that judges in other countries have fallen prey to the dangers of GAI, too.
One particular case, H. C. v. Contreras, No. F089316, 2026 WL 1493401 (Cal. Ct. App. Unpub. May 28, 2026), shows the negative result of an hallucinated citation. In that case, the California Appellate Court discussed and criticized the lower court’s reliance on a hallucinated citation when refusing to issue a protective order in a domestic child abuse case. The appellate court reversed, stating:
We have no difficulty concluding that it is an abuse of discretion for a court to rely in material part on fictional case authorities in rendering a decision or making an order. Reliance on fake cases is fundamentally incompatible with an informed exercise of discretion controlled by genuine principles of law. It seriously undermines the integrity of the outcome and erodes public confidence in our judicial system. It can also hinder meaningful appellate review.” (Campos and Munoz, supra, 118 Cal.App.5th at pp. 1123-1124.)
“This is not to say that the factors the court relied on were legally improper or that it could not have reached the same result applying proper legal reasoning grounded in genuine law. We hold only that the path the court took to reach its result constituted an abuse of discretion because it relied in material part on the holdings of nonexistent legal authorities” and misread the law. (Campos and Munoz…
Id. at *4 (Cal. Ct. App. May 28, 2026).
When a judge includes a hallucinated citation or hallucinated law in a decision the results can be far worse that an attorney including the same in a pleading. It may be some time before the hallucination is uncovered—if ever. Until it is discovered and corrected, the erroneous decision may have real world, practical effects far beyond the case in which it is filed.
There may come a time when judges will be sanctioned for incorporating hallucinated cases into official court documents, but that time does not appear to be approaching anytime soon. Instead, it appears that the judicial brethren are attempting to deal with the problem by a combination of public embarrassment and quite a bit of published advice. In terms of embarrassment, when a hallucinated citation is discovered in an opinion or court order, it is almost always withdrawn or appealed and quickly reversed. Of course, in these cases there are additional costs for the court and the litigants.
Short of banning judicial use of GAI, there are other ways to deal with the problems it creates. The National Center for State Courts has published several reports of judicial use of GAI incorporating results of surveys. One reported noted:
Although judges interviewed were keenly aware of the risks of GenAI, they nevertheless were finding value in GenAI tools to increase their productivity and efficiency. In addition to the use cases for GenAI discussed above, judges also reported using (or considering using) GenAI in the following ways:
- Assistance with drafting law review articles or research for law review articles
- Assistance with teaching classes at law schools
- Data analysis, such as when to hold hearings to maximize attendance
- Possible translation or transcription services in future if technology improves (though another judge said those were not good enough yet)
Obviously, some of these uses, such as writing for law reviews or preparing class materials may not have as many real-life negative aspects as using GAI with hallucinations. Yet uses that create legally authoritative documents are a whole different, and far riskier, category of use.
We suggest that because of the paucity of specific guidance for judges, they also might look to the types of workflow patterns that private law firms are developing. In essence, these firms have discovered that they cannot trust GAI generated documents. It is imperative that skilled and trained humans verify that every GAI document does not contain hallucinations or other errors and that these be corrected before the document is used.
As noted, GAI can be extremely helpful to sitting judges, especially those without adequate staffing. Even judges with clerks need warning. According to recent surveys, most law students use GAI all the time. Indeed, many would have difficulty if they were prohibited from using it by a judge. Still many, if not most, clerks have substantial workloads. It may well be necessary for judges either to decrease clerk workload or find other trained court personnel to do this critical task.
It is clearly time for lawyers, regulators, and the judiciary to formulate effective rules on judicial use of GAI to prevent further errors entering the stream of law and threatening the legitimacy of the judicial system.
READ THE FULL ISSUE OF LEMR, Vol. 7, No. 8
About Joseph, Hollander & Craft LLC
Joseph, Hollander & Craft is a mid-size law firm representing criminal defense, civil defense, personal injury, and family law clients throughout Kansas and Missouri. From our offices in Kansas City, Lawrence, Overland Park, Topeka and Wichita, our team of 25 attorneys covers a lot of ground, both geographically and professionally.
We defend against life-changing criminal prosecutions. We protect children and property in divorce cases. We pursue relief for clients who have suffered catastrophic injuries or the death of a loved one due to the negligence of others. We fight allegations of professional misconduct against medical and legal practitioners, accountants, real estate agents, and others.
When your business, freedom, property, or career is at stake, you want the attorney standing beside you to be skilled, prepared, and relentless — Ready for Anything, come what may. At JHC, we pride ourselves on offering outstanding legal counsel and representation with the personal attention and professionalism our clients deserve. Learn more about our attorneys and their areas of practice, and locate a JHC office near you.
