Loading case…
RECOMMENDED FOR PUBLICATION Pursuant to Sixth Circuit I.O.P. 32.1(b)
File Name: 26a0105p.06
UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT
No. 25-5623
Appeal from the United States District Court for the Eastern District of Kentucky at Covington. No. 2:24-cr-00028-1 — David L. Bunning, District Judge.
Decided and Filed: April 3, 2026
Before: CLAY, GIBBONS, and HERMANDORFER, Circuit Judges.
COUNSEL
Steven N. Howe, STEVEN N. HOWE, P.S.C., Williamstown, Kentucky, for Appellant.
PER CURIAM. This appeal involves John C. Farris's challenge to his sentence of imprisonment. But today’s opinion does not resolve the underlying merits of Farris’s appeal. We instead address the conduct of Farris's attorney, Steven N. Howe, who was appointed to represent Farris pursuant to the Criminal Justice Act, 18 U.S.C. § 3006A. By his own admission, Howe used artificial intelligence to draft the briefs in this case and then filed them without properly verifying the cited legal authorities. The result was multiple misrepresentations of law to this Court. This opinion details Howe’s conduct and the consequences that follow.
Page 2The U.S. District Court for the Eastern District of Kentucky appointed Howe to represent Farris in his criminal proceedings. Farris pled guilty to the crimes charged. At sentencing, the district court determined that Farris qualified for a two-level enhancement under § 3B1.1(c) of the Sentencing Guidelines for his leadership role during drug-trafficking activities. After Farris appealed that enhancement, our Court authorized Howe to continue serving as Farris’s attorney.
Howe filed two briefs — a principal brief and a reply — on behalf of Farris. Upon our initial review of the case, we began to suspect that Howe’s briefs were generated, at least in part, by artificial-intelligence software. The first tell was the file name of the principal brief: “CoCounsel Skill Results.” CoCounsel is the name of Westlaw’s internal artificial -intelligence platform. See Thomson Reuters Launches CoCounsel Legal: Transforming Legal Work with Agentic AI and Deep Research , Thomson Reuters (Aug. 5, 2025), https://perma.cc/ZU4ASWK6. From our vantage point, that filename abnormality suggested that Howe’s brief might have derived not from Howe's independent work, but directly from artificial -intelligence software.
Further suspicions arose when, during our substantive review of the briefs, we discovered three problematic citations:
Page 4 of the principal brief states, 'The Guidelines' commentary makes clear that '[m]ere presence or knowledge of the offense is not sufficient to make a person a participant.’ U.S.S.G. § 3B1.1 cmt. n.1.”
Page 10 of the principal brief states, 'The Sixth Circuit has reversed role enhancements on similar facts. In Washington , the Court held that 'simply facilitating the offense without exercising decision-making authority is insufficient.’ 715 F.3d at 985.”
Page 10 of the principal brief states, “Likewise, in Anthony , the Court vacated a § 3B1.1 enhancement because ‘[t]here was no evidence [the defendant] directed or supervised anyone else.’”
The reply brief repeats the latter two quotations. Each of these citations references genuine legal authorities. But the purported direct quotations do not appear in their cited sources. And upon deeper review, we were unable to locate any relevant legal authority that contained the same or Page 3substantially similar language as the above quotations. So, it did not appear that the misattributions involved mere citation mix-ups or transcription errors.
Moreover, the briefs Howe filed misrepresent the holdings of both United States v. Washington , 715 F.3d 975 (6th Cir. 2013), and United States v. Anthony , 280 F.3d 694 (6th Cir. 2002). In Washington , this Court upheld an enhancement under § 3B1.1 — that enhancement was not reversed, as Howe's principal brief asserts. 715 F.3d at 98384. And although the Court did vacate a § 3B1.1(a) enhancement in Anthony , it did so narrowly based on the proper counting methodology applicable to that enhancement —something irrelevant to Farris’s appeal. 280 F.3d at 700 –01. Indeed, contrary to Howe’s briefs, the defendant in Anthony conceded his role as a director and supervisor. Id. at 698.
Based on those discoveries, we directed the Clerk to issue a show-cause order on February 23, 2026. That order required Howe to provide a copy of each authority he cited in his briefs from either Westlaw or LexisNexis and discuss any discrepancies between the real authorities and his briefing. We further required Howe to “explain (1) who wrote the briefs, (2) whether generative artificial intelligence was used in the drafting of these briefs, (3) the processes that were used to cite-check each brief, and (4) whether and to what extent Howe utilized generative artificial intelligence in any district court filings in this case.” Order to Show Cause at 3.
Howe filed a timely response. In it, Howe admits that he used artificial intelligence to prepare both briefs he filed. According to Howe, he directed an unnamed 'staff' member to upload district court documents to Westlaw’s CoCounsel program to create a first draft of the principal brief. Howe Show Cause Response at 2. He then worked in that same file for six hours to supplement the draft produced by artificial intelligence. Howe notes that he repeated that same process for the reply brief.
By way of attempted explanation, Howe claims that this appeal was his first time utilizing Westlaw CoCounsel “in this way for a Court of Appeals brief.” Id. And he says that he was otherwise unfamiliar with the program. Howe's response states that his law office first acquired Westlaw CoCounsel in August 2025 — after the district court proceedings Page 4concluded -and that no artificial-intelligence software was used to prepare documents before that court. Howe notes that he has never been disciplined over his 40-year career, whether for improper use of artificial-intelligence software or otherwise.
Howe agrees that the briefs he filed before this Court contain legally erroneous content that was generated by artificial intelligence. He concedes that the three inaccurate quotations identified above were the product of artificial intelligence, that they do not appear in any legal authorities, and that his briefs misrepresented the holdings of both Washington and Anthony . Howe admits that those errors occurred because he failed to adequately review and verify the draft brief produced by artificial intelligence, and he accepts full responsibility for that error.
Howe's is the latest artificial -intelligence misstep our Court has confronted. See, e.g. , Whiting v. City of Athens , --- F.4th ----, 2026 WL 710568, at *4 – 5 (6th Cir. 2026). We therefore write to reiterate attorneys' baseline ethical obligations as they relate to the use of artificial intelligence.
New technologies present significant promise for the legal field. But all in the legal profession must be clear eyed about technology’s potential pitfalls. That mandate is especially critical in today's rapidly evolving artificial -intelligence landscape. Cf. Model Rules of Pro. Conduct. r. 1.1 cmt. 8 (A.B.A. 2012) (attorneys’ duty of competence requires keeping up with “changes in the law and its practice” including “relevant technology”). Howe claimed that he was “not familiar” with the CoCounsel program and did not scrutinize its incorporation into the briefing process. Howe Show Cause Response at 2. The risks of reflexively relying on artificial intelligence in the practice of law, however, are well documented. See generally A.B.A. Comm. On Ethics & Pro. Resp., Formal Op. 512 (2024) (gathering authorities). Attorneys should not utilize technology without knowing the ways in which it can be misused or contribute to inaccuracies. That remains true even when new tools are sponsored by trusted legal technology providers. See Varun Magesh et al., Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools , J. of Empirical Legal Stud. 9 – 16 (2025).
Page 5Further, attorneys who choose to use artificial-intelligence tools must do so in a manner consistent with their ethical obligations. See A.B.A. Task Force on L. & A.I., Addressing the Legal Challenges of AI: Year 2 Report on the Impact of AI on the Practice of Law , 47 – 48 (2025), https://perma.cc/4JES-DGMU (compiling state bar ethics rules and guidance on generative artificial intelligence). We do not purport to exhaust the full scope of ethical considerations here. But relevant steps may include reviewing and validating content produced by artificial intelligence; considering whether to disclose the use of artificial intelligence to clients or obtain informed consent; safeguarding confidential client information and preserving attorney-client privilege; implementing firm-wide policies governing the use of artificial intelligence; adhering to ethical billing practices when using artificial-intelligence tools; and keeping current with jurisdiction-specific guidelines.
New technologies, moreover, are no substitute for tried-and-true safeguards managed by practicing attorneys. Attorneys have an ethical obligation to verify the citations and propositions they submit to courts; that obligation reflects duties of competence and candor that apply no matter the tools attorneys use. See, e.g. , McCoy v. Ct. of Appeals of Wis., Dist. 1 , 486 U.S. 429, 440 – 41 (1988); Fletcher v. Experian Info. Sols., Inc. , 168 F.4th 231, 239 – 40 (5th Cir. 2026). So, attorneys who rely on artificial intelligence must remain diligent in supervising their work product and carefully examine the accuracy of every citation they present to this Court. Here, Howe’s reliance on “staff”— rather than himself or another attorney — to supervise the artificialintelligence-generated work product fell short of his obligations as attorney of record. See Model Rules of Pro. Conduct. r. 5.3 (A.B.A. 2012); Ky. Sup. Ct. R. 3.130(5.3) (2022).
That Howe’s briefs cited real legal authorities—as opposed to “hallucinations” featuring fictitious cases — does not absolve him. See Sanders v. United States , 176 Fed. Cl. 163, 169 & n.8 (2025) (collecting cases with invented authorities). Howe's failure to verify the artificial-intelligence output still resulted in the submission of false quotations and misleading legal arguments to this Court. Again, attorneys’ professional duties demand more.
We appreciate Howe’s timely response to the Court’s show -cause order as well as his candor in acknowledging his improper use of artificial intelligence. And we take note that this appears to be the first time a court has had occasion to address Howe's misconduct in his Page 6practice of law. But the fact remains that Howe committed inexcusable transgressions during the appellate phase of this case. And that misconduct had consequences. Among other things, it necessitated a significant use of judicial resources to investigate the suspected artificial-intelligence improprieties, coordinate a response, and facilitate additional steps of these appellate proceedings. As we order below, Howe's misconduct also warrants appointment of different appellate counsel to file new briefs -further delaying resolution of Farris's criminal appeal. That Howe was serving as court-appointed counsel for an indigent defendant through a publicly funded program only compounds the harm to our system of justice.
Based on the above, we conclude that the following measures are appropriate:
Additionally, by separate order issued on this same date, we remove Howe from further representation of Farris and order the Clerk to (i) appoint replacement counsel under the Criminal Justice Act, (ii) lock the briefs Howe filed on behalf of Farris, and (iii) reset the briefing schedule upon appointment of replacement counsel.