United States v. Farris
U.S. Court of Appeals for the Sixth Circuit · 6th Cir. · Kentucky bar guidance , Michigan bar guidance , Ohio bar guidance , Tennessee bar guidance
Verified April 26, 2026
- Citation
- United States v. Farris, No. 25-5623 (6th Cir. Apr. 3, 2026) (per curiam)
- Decided
- April 3, 2026
Summary
Steven N. Howe, court-appointed CJA counsel for defendant John C. Farris on appeal from a sentencing enhancement under U.S.S.G. § 3B1.1, filed principal and reply briefs containing fabricated quotations and misrepresented holdings. A staff member had used Westlaw CoCounsel to generate the drafts; Howe did not verify the output. The panel's suspicions were first triggered by the principal brief's file name, "CoCounsel Skill Results." On review, the panel (Clay, Gibbons, and Hermandorfer, JJ., per curiam) identified problematic citations whose purported direct quotations did not appear in the cited sources, including commentary attributed to U.S.S.G. § 3B1.1 cmt. n.1, and found the briefs misrepresented United States v. Washington, 715 F.3d 975 (6th Cir. 2013), and United States v. Anthony, 280 F.3d 694 (6th Cir. 2002).
- AI tool:
- Westlaw CoCounsel
What sanction did the court impose?
The panel imposed four sanctions: (1) denial of all CJA compensation for time spent on the appeal; (2) referral to the Chief Judge for potential disciplinary proceedings; (3) service of the published opinion on the district court and on state bar disciplinary authorities; and (4) removal of Howe from the representation, with appointment of replacement counsel and a reset briefing schedule. The opinion was recommended for publication under Sixth Circuit I.O.P. 32.1(b).
Why does United States v. Farris matter for law firms using AI?
Farris is a meaningful escalation in the AI-hallucination sanctions line. A federal appellate panel did not merely fine counsel: it stripped CJA compensation for the entire appeal, removed counsel mid-case, and pushed the matter to bar authorities, all in a published opinion. For managing partners, the case underscores that delegating generative-AI drafting to staff without a verification protocol is itself the sanctionable conduct, and that a legal-vertical tool branded for lawyers (Westlaw CoCounsel) is no safer than a general-purpose chatbot when output goes unverified.
Sources
Primary sources
Further reading
Source PDF is a Westlaw printout mirrored from the Damien Charlotin hallucination database. We are working to add the underlying court docket (PACER, CourtListener, or court website) as a second source.