Summary
In *United States v. Farris*, the Sixth Circuit paused the merits of a criminal appeal to address the conduct of appointed counsel, who admitted using Westlaw CoCounsel to draft briefs and then filing them without adequately verifying the authorities. The court found that the briefs contained false quotations and misleading characterizations of real cases, removed counsel from the appeal, denied compensation for the appellate work, and referred the matter for possible discipline.
Why It Matters
This is a strong direct legal misuse story because it turns abstract AI-risk talk into a concrete appellate record about what goes wrong in practice:
- legal AI can distort real authorities without inventing fake case names
- trusted legal vendors do not eliminate the lawyer's personal verification duty
- attorney supervision obligations apply to both staff and AI-assisted work product
- courts are willing to spend judicial resources investigating suspicious AI drafting artifacts and to impose real consequences
For lawyers operationalizing AI, the key lesson is not "never use AI," but "never outsource citation accuracy, legal characterization, or filing responsibility to it."
PI Tool Angle
`n/a`
What the Source Says
The opinion says the court first suspected AI use because the principal brief's file name was "CoCounsel Skill Results." It then identified three problematic citations, including quotations that did not appear in the cited authorities and arguments that misdescribed the holdings of *United States v. Washington* and *United States v. Anthony*. In response to a show-cause order, counsel admitted that a staff member uploaded district-court materials into CoCounsel to create a first draft, that he then worked from that AI-generated file, and that he failed to verify the output adequately. The Sixth Circuit ordered that counsel receive no Criminal Justice Act compensation for the appeal, referred the opinion for possible disciplinary action, and appointed replacement counsel with a new briefing schedule.