Judge Sanctions Both Sides for AI Usage

Last week, a federal judge in Mississippi sanctioned lawyers on both sides in a legal dispute for using generative AI in their filings. 

The story was first covered by Jason Koebler at 404 Media and has since been picked up by Gizmodo, The Cool Down, and the Mississippi Free Press.  

At the center of the case is a basic contract dispute. Lawyer Tom Withers sued the City of Aberdeen, Mississippi, for  unpaid legal fees. Withers, to be clear, did not represent himself in this case and was not part of the sanctions.

Instead, four attorneys, two for each side, were sanctioned by Northern District of Mississippi Sharion Aycock. The reason was that both sides of the cause used AI to generate findings and then failed to fact-check them. As a result, both sides had multiple “hallucinations,” namely fabricated legal arguments and sources.  

According to Aycock, “This case presents the Court with an unusual scenario—attorneys for both litigants engaged in similar sanctionable conduct.” All four attorneys were removed from the case and barred from appearing before the Northern District of Mississippi for two years. They were also ordered to pay a total of $8,000 in penalties.

This decision wipes out the legal team on both sides. Withers, who was represented by Kathleen Wilson and Shauncey Hunter Ridgeway, and the City of Aberdeen, which was represented by Mark McClinton and Kathryn Young Williams. Ridgeway and McClinton testified that they did not use AI but failed to check the work of their colleagues before signing off on them.

For Wilson, the story gets even worse. According to the judge, Williams has faced a similar sanction in March in the U.S. Bankruptcy Court for the Western District of Louisiana. Similarly, she admitted to submitting filings with AI hallucinations. Those filings came two months after she apologized in Aycock’s court.

Williams and Wilson were also serving as pro hac vice attorneys in the case; this means that they were licensed in a different state but were allowed to practice in Mississippi for this case. The judge indicated she may revoke that status, barring them from practicing in Mississippi in the future.

The judge found that both Ridgeway and McClinton were “negligent and careless but not purposeful” in their actions and appreciated their efforts to correct the issues. However, the sanctions still stand as they were both responsible for the actions of the out-of-state attorneys operating under their supervision.

It’s unclear what will happen with the actual case after this. 

To put it mildly, this case is a mess. However, it’s far from the only time that attorneys have been sanctioned for using AI.

A Growing Problem

As an April 2026 article by Martin Kaste at NPR notes, this has been an ongoing issue in the legal world. Some attorneys have rushed to embrace the use of AI, seeing it as a way to save time on writing, research, and other tasks.

However, this usage has repeatedly backfired on lawyers. An AI hallucination tracker by Damien Charlotin has tracked over 1,600 instances of lawyers being sanctioned or otherwise admonished for using AI. The list includes a mixture of pro se litigants, attorneys, and even some judges who have been flagged for using AI due to hallucinations citations.

What makes this latest story unique is that both sides of the case are accused of doing the exact same thing. While it’s an amazing (and in some ways humorous) coincidence, it is also a coincidence that was bound to happen at some point.

The other thing that makes this case unique is that the plaintiff is himself a lawyer. While he did the right thing to hire outside counsel, I suspect, right now, he would have been better off if he had gone pro se.

To be clear, that is a sentence that should never be written. I feel wrong even typing it. But his lawyers set his case back, and the only saving grace is that the other side did the same thing.

This case clearly illustrates that this is a growing problem for courts, and it’s one that they need to address immediately.

Addressing the Issue

If there’s one thing that Charlotin’s tracker shows, it’s that the responses from judges are all over the map.

For example, of the 1,135 cases tracked in the United States, only 167 involved monetary penalties. When penalties were issued, the bulk of them were in the $1,000 – $5,000 range, though some were over $50,000.

Even fewer were given a disciplinary referral. 

In short, the penalties for using AI and submitting false citations ranged from having to apologize and correct the document to a massive fine and a disciplinary referral.

To me, this is not an AI issue. The question is much more fundamental. Is an attorney responsible for ensuring that the work they submit is accurate? Yes, mistakes happen, and we need to allow grace for that. But the lawyer submitting the work has the final responsibility for it. This is whether they wrote it themselves, had a junior do it, or used AI

To be clear, this is equally true for pro se litigants. You have the same obligations regardless of what legal training you have had.

Lawyers not validating citations need to be taken seriously. The citations need to exist, they need to say what the filing says it does, and any quotes from it need to be accurate. This is fundamental stuff and core to the entire practice of law.

Using AI is not an excuse. If anything, it’s a reason to exercise even more care and diligence. But, as we’ve discussed before, exercising that diligence takes significant time and energy, which defeats much of the purpose of using AI in the first place.

If that’s the case, maybe we should just accept that AI is not ready to be used in this way and move on. If it doesn’t do better work and doesn’t save time, what is it really improving?

Bottom Line

If I were a lawyer and I were not using AI, I would be promoting that fact or at least making it more known. Even if I used it in a limited capacity, such as to aid in research, I would be promoting it through an AI disclosure that clearly states AI does not write my filings. 

In the end, what makes this case unique is not that lawyers were sanctioned for using AI, but that both sides of the case did it. It’s a coincidence, but perhaps not as amazing of one as it seems.

Stories like this one do active harm to the legal profession. As Charlotin’s tracker shows, over half the cases tracked involved pro se litigants. But why would a layperson think they can use AI to file a lawsuit? Because, according to the tracker, nearly all of the remaining cases were filed by lawyers. 

Every time that a lawyer is caught with AI hallucinations in their filings, it gives more people the impression that the human component of the profession is not necessary. After all, if lawyers are just generating their filings, why shouldn’t they?

Of course, everyone who has been anywhere close to the legal profession knows the answer to that question. But it’s not people in or near the legal profession who are being fooled. It’s those who don’t know the legal field and have bought into the AI hype.

There needs to be clearer rules and more meaningful consequences for submitting false citations, regardless of whether they were AI-generated or not. 

In this case, the judge did the right thing. Stiff sanctions and a penalty totaling $8,000 are a significant deterrent. Other courts need to follow this example.