By Jeanne M. Huey
Two recent Sixth Circuit Opinions drive home the point: Staying out of trouble, even when using new technology, is easy if lawyers do their job i.e. read and verify every citation before filing.
We have become used to a constant stream of headlines about all the ways that fake citations have ended up in court filings. But the real issue isn’t howthese citations came to be, it is that the filing was submitted to the court at all. Lawyers have long been required to verify every authority they cite before signing their name and filing something with the court. That duty hasn’t changed just because technology has. Two recent Sixth Circuit Court of Appeals decisions involving misuse of artificial intelligence (AI) underscore the point.
No Response? No Matter
In Whiting v.City of Athens, the U.S. Court of Appeals for the Sixth Circuit (Eastern District) found, in part, that the plaintiff’s filings “repeatedly misrepresented the record, cited non-existent cases, and cited cases for propositions of law that they did not even discuss, much less support.” The court issued a show-cause order and instructed the lawyers to address questions about AI use and their cite-verification practices, among other issues. The lawyers refused and instead challenged the order on several grounds, asserting both work-product and attorney-client privilege doctrines with respect to their AI use. The court rejected all of their arguments.
Citing multiple reasons for issuing harsh sanctions, the court determined that the appellate rule addressing frivolous appeals, Rule 38, was not “up to the task,” on its own, of imposing the sanctions that were warranted. The court then employed its inherent power, in addition to Rule 38, to sanction the lawyers, noting that: “Courts have these powers because when certain forms of misconduct are ‘not instantly suppressed and punished, demoralization of the court’s authority will follow.’” Citation omitted.
Finding that the lawyers’ conduct in this case had “brought the profession into disrepute” and “sullied the reputation of our bar,” the court imposed sanctions consisting of an award of attorney fees and a double award of costs to the opposing party as well as a monetary sanction to be paid to the court’s registry. The court also referred the lawyers to the chief judge to be considered for discipline. The opinion is worth reading just for the five-page appendix in which the court analyzes each of the offending citations it had identified.
Use of Paid Legal Research Tools Still Requires Verification
Just a month later, the same court had an opportunity to address AI use gone wrong in a criminal appeal. In United States v. Farris, the court suspected from the start that briefs filed by a lawyer appointed to represent a criminal defendant may not have been the “independent” work of the lawyer. “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.” Upon further review, the court found that the filings contained both false quotations and misleading legal arguments.
The lawyer admitted to using AI and filing the briefs without verifying the cited authorities. He explained to the court that he had “directed an unnamed ‘staff’ member” to use Westlaw’s CoCounsel tool to create a first draft before editing (but not verifying), signing, and filing the brief. The lawyer admitted that he was “not familiar” with the program and had not used it to draft briefs before. The court imposed sanctions and emphasized that “Attorneys should not utilize technology without knowing the ways in which it can be misused or contribute to inaccuracies.” The court ordered the removal of the attorney from the appointment, denied him compensation, and initiated disciplinary proceedings.
Verify the Law—or Answer to the Court
With these opinions, the Sixth Circuit drives home an obvious point that bears repeating only because it keeps being ignored. As the court stated in Whiting: “Citing even a single fake case can be sanctionable because ‘no brief, pleading, motion, or any other paper filed in any court should contain any citations—whether provided by generative AI or any other source—that’ a lawyer has not personally ‘read and verified,’” (citation omitted).
In other words, staying out of trouble, even when using new technology, is easy if lawyers read and verify every citation before filing.
A Refresher: The Relevant Rules
The rules have long required lawyers who sign court filings to confirm that the cited authority exists and supports their arguments. False citations in court filings almost certainly violate Model Rule 3.3 (candor to the tribunal), and Model Rule 1.1 (competence), as failing to verify, or fabricating, authority falls short of the standard expected of a competent lawyer. If technology was used to generate and/or find the authorities, it would also be a breach of Rule 1.1 Comment 8, which requires lawyers to be competent in the technology we use in practice. If particularly egregious, citing to bad law may violate Model Rule 8.4’s prohibition on misrepresentation. Supervisory failures under Model Rules 5.1 and 5.3 can also give rise to discipline, even if the supervisor did not sign the filing.
And, of course, inaccurate or misleading citations likely violate Federal Rule of Civil Procedure 11(b)(2) or its equivalent in state court. If the misrepresentations or false statements prolong proceedings, sanctions in the trial court may be handed down under 28 U.S.C. § 1927. A frivolous appeal can trigger sanctions under Federal Rule of Appellate Procedure 38. Finally, all courts can use their inherent authority to issue sanctions to protect the integrity, efficiency, and legitimacy of the judicial system.
© 2026 by the American Bar Association. Reproduced with permission. All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or stored in an electronic database or retrieval system without the express written consent of the American Bar Association.
Legal Ethics Today 