AI Regulation Tracker  /  Courts and litigation

Sixth Circuit Fines Two Attorneys $15,000 Each Over Fake Citations, With Full Fees, Double Costs, and a Discipline Referral

A new federal appeals court has joined the AI-citation sanctions map, and it set a high-water mark. On March 13, 2026, the Sixth Circuit sanctioned two Tennessee attorneys for filing appellate briefs riddled with more than two dozen fake or misrepresented citations. It fined each lawyer $15,000, ordered them to pay the opposing side's full appellate fees plus double costs, and referred them for disciplinary proceedings.

By Anthony Guerriero  ·  The Leveraged Years Regulatory Research  ·  Published July 20, 2026  ·  Last updated July 20, 2026

Another federal appeals court has now weighed in on lawyers filing briefs full of citations that do not exist, and the Sixth Circuit did not do it quietly. On March 13, 2026, in Whiting v. City of Athens, a three-judge panel sanctioned two Tennessee attorneys, Van Irion and Russ Egli, for consolidated appellate briefs that the court found were shot through with fabricated and misrepresented authority. This is not a trial court warning shot. It is a published circuit opinion, and it stacks penalties in a way that makes it the toughest treatment of fake citations at the appellate level to date.

What pushes this case onto the map is the combination of severity and reach. The panel did not pick one sanction. It imposed a $15,000 punitive fine on each attorney, made them jointly responsible for the opposing party's full attorney fees on the appeal, added double costs, and then referred both lawyers to the chief judge for formal disciplinary proceedings under the circuit's local rules. The court explained the size of the fine plainly: the misconduct spanned multiple cases, both attorneys had prior discipline for lack of candor, and, in the court's words, smaller fines have been inadequate.

What the briefs actually contained

The court's description of the record is the part every litigator should sit with. Across the consolidated appeals, the panel counted what it called a conservative estimate of more than two dozen problem citations. Some were citations to cases that simply do not exist. Others were real reporter citations that pointed to entirely unrelated decisions, including two Tenth Circuit cases that had nothing to do with the fees and sanctions issues the brief cited them for. Others were quotations that did not appear anywhere in the authority quoted. This was not one stray hallucination that slipped past a busy associate. It was a pattern dense enough that the court treated it as a fundamental breach of the duty of candor.

The panel was blunt about what a lawyer is supposed to notice. It wrote that "any reasonable attorney should know that a case is meritless if the only authority on which he can rely is a figment of imagination." That line matters because it forecloses the most common excuse. The problem is not only that fabricated cases are fake. It is that a competent lawyer reading his own brief should recognize when the only support for an argument is invented, whatever produced it.

The AI question the court refused to make the crux

The show-cause process here probed whether generative AI was behind the fabrications, and this is where the opinion is most useful. The court did not rest its ruling on a finding that AI generated the fake cites. It deliberately wrote a rule that does not depend on the tool at all. In the panel's words, no filing "should contain any citations, whether provided by generative AI or any other source," that the lawyer has not personally "read and verified."

Read that carefully, because it is the whole point. The duty the Sixth Circuit is enforcing is not a duty to avoid AI. It is a duty to personally read and verify every citation before it goes in a brief, and that duty is identical whether the cite came from a chatbot, a junior associate, a paralegal, an old memo, or the lawyer's own memory. Framing the rule this way is smart, and it is bad news for anyone hoping to argue that the AI was to blame. The court has removed the tool from the equation. The signature on the brief is the verification, or it should be.

Where this sits on the sanctions map

This is now part of a clear pattern across the federal circuits. Trial courts have been sanctioning lawyers for AI-fabricated citations for a while, and appellate courts have started to follow. What makes Whiting stand out is the ceiling it sets. The combination of a five-figure punitive fine per lawyer, full fee-shifting, double costs, and a disciplinary referral, all in one order, is more aggressive than most of what has come before at the circuit level. When a court reaches for every penalty it has and says on the record that smaller fines have not worked, it is sending a message to the bar, not just to two lawyers in Tennessee.

The prior-discipline detail is worth holding onto as well. Both attorneys had records that fed the court's sense that a modest sanction would not deter. One had a public censure for lack of candor; the other had a multi-year suspension in a district court for dishonesty. Courts are watching for repeat conduct, and they are willing to escalate hard when they see it. A first offender with a clean record and a fast, candid correction is in a very different position than a lawyer with a history, which is exactly why the response to a discovered error matters so much.

What US litigators and firm managers should do now

The operational takeaway is not complicated, and it is not optional. Treat every citation in every filing as something you personally read and verified, and build a process that proves you did. Pull the actual case for every cite. Confirm the case exists, that it says what your brief claims, and that any quotation appears verbatim in the opinion. Do this whether the draft came out of a generative AI tool, a research memo, a form brief, or your own recollection, because the court will not care where it came from. If you supervise other lawyers, the verification duty runs to the name on the signature block, so a partner cannot outsource the risk to an associate or a tool.

For anyone using AI in a litigation practice, the honest framing is that these tools are fine for drafting and terrible as a citation source you trust without checking. The lawyers who get hurt are not the ones who used AI. They are the ones who filed what the AI produced without reading it. The discipline referral in this case is the reminder that the exposure is not only financial. A pattern of fabricated citations can put your license in front of the people who decide whether you keep it. Verify first, every time, and the tool question takes care of itself.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any court order or professional-responsibility duty applies to your situation with qualified counsel in the relevant jurisdiction.