AI Regulation Tracker / Courts and professional responsibility
Seventh Circuit Admonishes Attorney for AI-Style Fake Citations but Declines to Sanction
In Dec v. Mullin, decided March 30, 2026, the Seventh Circuit found two non-existent case citations and a false quotation in an appellate brief, called them out in a published opinion, and then imposed no penalty beyond a public admonishment because the lawyer owned the mistake and was contrite. It is the lenient end of a widening circuit split over how hard to hit lawyers for hallucinated authority. The duty to check every citation and quote did not move.
Set the immigration dispute aside for a moment, because the part of this decision that matters to working professionals is the last few pages. Danuta Dec lost her appeal on jurisdiction, and the Seventh Circuit affirmed. But the court flagged a second problem in her lawyer's opening brief. As the opinion puts it, the brief "cites two cases that do not exist and recounts an imaginary quotation," and those errors "have tell-tale signs of AI hallucinations, in which an AI large-language model generates a fictional or inaccurate output." The court did not let it pass. It devoted a full section of a published opinion to the conduct.
What happened next is the interesting part. When asked at oral argument, the lawyer could not explain the citations and denied using AI. In a later letter she doubled down, saying she is "way too technologically challenged to even attempt to use AI," and that she had "presumably copied and pasted that portion from another brief several months ago" but could not locate it. She took responsibility in plain terms. The court quotes her: "This was clearly my error." That posture is why she walked away with a warning instead of a bill.
What the court actually held
The Seventh Circuit stated the duty in language that leaves no wiggle room. "Whatever AI's future in litigation," the court wrote, "our concern lies with trained lawyers failing to check the accuracy of legal citations and quotations in their filings." It went further and pointed at the tools available, noting that entire briefs can now be run through Westlaw or LexisNexis to check citations, so there is no good excuse for filing fake ones. The failure the court cared about was not using AI. It was signing a document without verifying what was in it.
Then the court chose its remedy carefully. It admonished the lawyer, and it explained why it stopped there. In its words: "Beyond this public admonishment, we decline to issue an order for Dec's counsel to show cause why she should not receive any further sanctions. We are persuaded by her acceptance of responsibility and sincere apology that these false citations were not made knowingly or intentionally." The court added that the errors were few and mainly supported an undisputed standard of review, so nobody was misled on a contested point. The fact that opposing counsel also missed the fake citations gave the court further pause, though it said that mattered to a lesser degree.
Why this is the lenient end of a split
Read this decision next to what other circuits are doing and the picture sharpens. Courts around the country are now regularly catching hallucinated authority in briefs, and they are not all responding the same way. Some have imposed monetary sanctions, ordered lawyers to notify clients, or referred them for discipline. The Seventh Circuit, on these facts, chose the softest tool it has, a public admonishment in the case reporter, and expressly declined to make counsel show cause for anything worse. That is a real data point about where the discretion band sits when a lawyer is honest and contrite.
But do not confuse a lenient sanction with a lenient rule. The Seventh Circuit was explicit that under its standards for professional conduct, lawyers promise they will "not knowingly misrepresent, mischaracterize, misquote, or miscite facts or authorities in any oral or written communication to the court," and that every filing represents that the legal contentions are warranted. The court called adherence to those rules "crucial to the administration of justice" and spelled out the cost of violations: the court has to track down the false citations, seek explanations, and evaluate the damage, which "generates unnecessary work on a case within a case." The leniency was about this lawyer's candor. The obligation was stated as absolute.
What this means for US professionals
If you are a lawyer in Illinois, Indiana, or Wisconsin, this opinion is now on your shelf, and it tells you two things at once. First, the court will read your citations, and it will notice when they are invented, whether or not you used a chatbot. Second, if you do slip, how you respond will drive the outcome more than the slip itself. The lawyer here got the lightest possible treatment because she took ownership immediately and completely. A lawyer who stonewalls, blames the tool, or minimizes is not going to land in the same place, and other circuits have already shown they will reach for fines and referrals.
The broader lesson runs past the bar. Any professional who signs, files, or submits work is now operating in an environment where fabricated content can enter a document through an AI tool, a borrowed template, or a rushed cut-and-paste, and where the reviewer may catch it and put your name on the problem in public. The Seventh Circuit's framing is the useful one to carry into any field: the tool does not own the duty, you do. Verify before you sign. Run the check that takes ten minutes so you never have to explain a fabrication that takes ten pages. The efficiency of AI is real, and it changes none of that.
Questions professionals are asking
What did the lawyer actually do wrong in Dec v. Mullin?
Her opening brief cited two cases that do not exist and included a quotation that was never written, all in the standard of review section. The court said these had the tell-tale signs of AI hallucinations. She denied using AI and said she likely copied the passage from another brief she could not locate.
Why did the court impose only an admonishment?
The Seventh Circuit was persuaded that the false citations were not made knowingly or intentionally, because the lawyer accepted responsibility and apologized sincerely. The errors were also few and supported an undisputed legal standard, so no one was misled on a contested point. The court declined to order her to show cause for any further sanction.
Does this mean AI hallucinations are treated leniently now?
No. Dec sits at the lenient end of a split among the circuits. Other courts have imposed monetary sanctions, client-notification orders, and disciplinary referrals for the same conduct. What earned leniency here was the lawyer's candor and contrition, not a softer standard. The duty to verify citations and quotations was stated as absolute.
Did the court say the lawyer was required to avoid AI?
No. The court's concern was not the tool. It was trained lawyers failing to check the accuracy of citations and quotations in their filings. The court noted that entire briefs can now be checked through Westlaw or LexisNexis, so the failure was in not verifying, whether or not AI was involved.
What should lawyers and other professionals take from this?
Verify everything you file or sign, because the reviewer may catch a fabrication and put your name on it in a public record. If an error slips through, own it fully and promptly, since how you respond will shape the outcome more than the error itself. The efficiency of AI does not shift the duty to the tool. It stays with you.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any ruling or professional-conduct rule applies to your situation with qualified counsel in the relevant jurisdiction.