All Rise for the Honorable Judge Claude: Responding to the Use of AI in the Federal Judiciary

On August 31, 2026, the Fifth Circuit held oral arguments regarding the use of artificial intelligence (“AI”) in the courtroom.[1] Of course, this is not the first time that a court has had to address the usage of the 21st-century boogeyman in the practice of law, as the number of attorneys who are facing sanctions for submitting AI-written filings grows by the day.[2] However, this case is far different from the other cautionary tales of AI that have permeated the legal field over the past few years. Instead, the Fifth Circuit is confronting a far more novel and unsettling usage of technology in the halls of justice: the use of AI by a federal district judge. 

United States District Judge Henry Wingate of the Southern District of Mississippi has come under fire for an injunctive order he issued in July of 2025.[3] The order, which Judge Wingate admitted was written using AI, contained significant errors, including referencing parties that were not in the case and law that did not exist.[4] The Fifth Circuit now faces the decision of whether Judge Wingate should be reassigned from the case.[5]

This is not the only time that a federal judge has admitted to using AI to draft error-riddled orders. Judge Julien Xavier Neals of the District of New Jersey previously admitted to releasing AI-written orders containing substantial errors.[6] Both Judge Wingate and Judge Neal stated that members of their chambers used AI in some part of the order-writing process.[7]

Unregulated AI use, when utilized irresponsibly, poses a significant threat to the integrity of the judicial system and public policy at large. For instance, a December 2024 Gallup poll showed that only 35% of Americans had confidence in the Nation’s judicial system.[8] Meanwhile, a 2026 poll by the Pew Research Center found that about 32% of Americans believe AI will negatively affect the criminal justice system, while just 19% believe it will have a positive effect.[9] That same Pew Research Center study showed that 50% of respondents were “more concerned than excited” about the increasing role of AI in daily life.[10] The data is clear that a large portion of Americans are, at the very least, skeptical about the future of AI, with a significant subgroup pointing that skepticism towards the United States’ judicial system.[11] With faith in the Nation’s judicial system already at a concerning low, steps must be taken to ensure that the misuse of AI does not further shake the loose ground that the judicial system currently sits atop.[12]

There is also a potential constitutional question underpinning the usage of AI in the judicial system. While the Constitution does not outright guarantee a human judge to parties in a trial, such a rule can certainly be implied by the spirit and language of the text.[13] For instance, Article III of the Constitution states that federal judges “shall hold their Offices during good Behaviour.”[14] The issue here is that AI is not capable of “good Behaviour”—it is a complex, unliving technology without a moral compass or the ability to form intent. While this phrase seems to close the door to the potential of fully AI judges, it should also limit the scope of what federal judges can use AI for. A judge should not elevate AI into a position of making legal decisions, as this would essentially replace the judge with an entity that is incapable of serving as a judge. One would hope that no judge would intentionally use AI to make legal decisions. However, the error-riddled orders of Judge Wingate and Judge Neal are enough to warrant swift action by Congress and the Judicial Branch.

As for the resolution of the case of Judge Wingate and the future of AI in the courtroom, the Fifth Circuit is likely within its power to reassign Judge Wingate from the case and overrule the AI-written order. Under 28 U.S.C § 2106, the Supreme Court and all appellate courts of the United States may vacate any order of a court and “may remand the cause and direct the entry of such appropriate judgment, decree, or order, or require such further proceedings to be had as may be just under the circumstances.”[15] This authority includes the power to reassign a judge from a case.[16]

While no case law currently exists regarding reassignment for the use of AI, the Fifth Circuit has two tests for determining when to reassign a judge from a case.[17] The first test, the “stringent” test, considers three factors: The first prong examines whether the judge would be reasonably expected to have substantial difficulty ignoring their previously expressed views or erroneous findings.[18] The second prong looks at whether reassignment is needed to “preserve the appearance of justice.[19] The final prong inquires into whether reassignment would be wasteful in comparison to the gains made by the preservation of the appearance of justice.[20]      

     The “lenient” test asks whether “an objective observer” would “question the judge’s partiality.”[21] Under these tests, it is unclear whether the Fifth Circuit would actually exercise its power of reassignment against Judge Wingate. On one hand, Judge Wingate made a particularly egregious mistake, which could warrant reassignment to “preserve the appearance of justice.”[22] However, these tests seem more focused on cases in which the judge’s partiality is called into question, as opposed to cases in which the judge makes a mistake.      

Either way, there is certainly a moral argument at play here as well. Since this is the first instance of a judge facing reassignment for the misuse of AI, the Fifth Circuit bears the burden of setting non-binding precedent for the rest of the Nation to follow. It can either choose to exercise the full might of its powers of reassignment and make it clear that the misuse of AI by a judge is a threat to justice, or it can choose to forgive and forget and let the moment pass it by.

Although the Fifth Circuit faces the heavy burden of whether to reassign Judge Wingate, there is little that the Fifth Circuit can do in this case to deter future cases from being plagued with AI-hallucinations. In fact, there is little anyone can do about it. The District Court of Nevada recently ruled that judicial immunity extends to when a judge uses AI to write an order, meaning that a citizen could not sue for judicial misuse of AI.[23] Likewise, even Congress can only do so much to punish a judge who misuses AI, short of impeachment.[24]

However, what Congress and the Judiciary can do is set stricter guidelines regarding the use of AI. Whether they rely on means such as the Federal Rules of Civil Procedure, statutes, or local rules, it is imperative that Congress and the Federal Judiciary take swift action to prevent another AI-plagued order from coming out of the chambers of a judge again.


[3] Scarcella, supra note 1. 

[4] Id.

[5] Id. 

[7] Id. 

[9]  Michelle Faverio & Emma Kikuchi, Key findings about how Americans view artificial intelligence, Pew Rsch. Ctr. (March 12, 2026), https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence.

[10] Id. 

[11] Id.

[12] Id.

[13] See U.S. Const. amend. V (guaranteeing the right to due process, but making no reference to the type of judge needed to facilitate due process); See also U.S. Const., amend. VI (guaranteeing the right to a jury of one’s peers, which implies human jurors, but making no similar reference with regard to judges).

[14] U.S. Const. art. III, §1.

[15] 28 U.S.C § 2106.

[16] Liteky v. U.S., 510 U.S. 540, 554 (1994).

[17] S.E.C. v. Barton, 135 F.4th 206, 228 (5th Cir. 2025)

[18] Id.

[19] Id.

[20] Id.

[21] Id.

[22] Id.

[23] Phillips v. Parlade, 2026 WL 2364528 (D. Nev. Aug. 12, 2026).

[24] U.S. Const. art. II, §4.