By this point, I think all of us have seen stories about lawyers who have gotten in hot water because they relied on AI to come up with legal authority, only for the AI-generated citations turning out to be phony. In the following guest post, Ommid C. Farashahi and Melissa Y. Gandhi of the BatesCarey LLP law firm take a look at this recent phenomenon and consider the ethical and profesional implications. I would like to thank Ommid and Melissa for allowing me to publish their article as a guest post on this site. I welcome guest post submissions from responsible authors on topics of interest to this site’s readers. Please contact me directly if you would like to submit a guest post. Here is the author’s article.
***********************
Most professionals use artificial intelligence, or AI, for one purpose or another. For attorneys, whose profession depends heavily on research and writing, the appeal is obvious. But no matter how great it is, AI will not protect us from the fallout from misuse, and it will not absolve us from the ethical duties we owe to our clients.
One of the most cited cases in this space is Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), in which the court sanctioned attorneys $5,000 for filing a brief that contained non-existent judicial opinions and fabricated quotes and citations generated by AI. The attorneys were also directed to provide their client, as well as each judge falsely identified as the author of the fabricated cases, with a copy of the hearing transcript and the Court’s Opinion and Order imposing sanctions. The Court concluded that the penalty was “sufficient but not more than necessary to advance specific and general deterrence for acting in bad faith.” Similar sanctions and sentiments have been made by other courts. However, as violations increase in frequency, it seems as if courts are taking a stronger position.
Just last month, in Frankie Johnson v. Jefferson S. Dunn, et al., No. 2:21-cv-01701-AMM (N.D. Ala. July 23, 2025), the Court issued a Sanctions Order against attorneys who admitted to filing motions with fabricated legal authority generated by AI. In her 51‑page sanctions order, Judge Anna Manasco emphasized the serious nature of fabricating legal authority, stating that it “demands substantially greater accountability than the reprimands and modest fines that have become common as courts confront this form of AI misuse.” The Court held that an appropriate and reasonable sanction must: (1) have sufficient deterrent force to make misuse of AI unprofitable for lawyers and litigants; (2) correspond to the extreme dereliction of professional responsibility that sham citations reflect (whether generated by artificial or human intelligence); and (3) effectively communicate that made-up authorities have no place in a court of law.
In Johnson, the “punishment that fit the crime,” so to speak, was:
- Public reprimands for the attorneys’ misconduct, requiring them to provide a copy of the Court’s Order to their clients, opposing counsel, and the presiding judge in every pending state or federal case in which they are counsel of record, as well as to every attorney in their law firm. Compliance ordered to occur within 10 days of the order, and certification of compliance filed within 24 hours thereafter;
- In order to “further effectuate the reprimands and deter similar misconduct,” a directive to the Clerk of Court to submit the Order for publication in the Federal Supplement;
- Disqualification of the attorneys from further participation in the underlying matter;
- Directive that the attorneys, within 24-hours, provide the Clerk of Court with a listing of jurisdictions in which they are licensed to practice law; and
- Directive for the Clerk of Court to serve a copy of the Order to the General Counsel of the Alabama State Bar and any other applicable licensing authorities for further proceedings as appropriate.
Of note, the sanctions above were imposed only on the culpable attorneys, not their firm, mainly because the firm had AI-related policies, including an AI committee and protocols discouraging reliance on generative AI without verification. The Court found that sanctions against the firm weren’t warranted because it had “acted reasonably in its efforts to prevent this misconduct and doubled down on its precautionary and responsive measures when its nightmare scenario unfolded.”
Our main takeaway from these cases is that courts may accept that AI has changed how attorneys work (with emphasis seemingly being placed on the misuse of AI rather than the tool itself), but they will not accept a change in our professional obligations. Further, even in situations where attorneys are sanctioned for AI-related misconduct, their firm may be insulated if they have instituted reasonable protocols and procedures