Select Page
artificial intelligence 4111582 1920
AI Slop Or Novel Arguments? A Texas Judge Goes Off The Rails 3

Hallucinations in briefs and court filings. Citations to cases that don’t exist. Citations to cases for propositions that are incorrect or that make little sense. It’s happened over 2,000 times according to Damien Charlotin’s database. And when lawyers do it, they deserve to be called out. That’s old news. 

But what happens when lawyers use AI to write a brief that reads like AI slop? Or make arguments so nonsensical that you wonder whether they came from AI rather than a human being?

Well, that’s exactly what happened recently in a federal district court case in Texas, as reported in a Business Insider article. 

The Slop Order

U.S. District Judge Mark Pittman ripped three lawyers to shreds not for hallucinated cases but for filing motions containing what he described as blatant AI slop, according to the article. 

The judge concluded the arguments were so weak and legally flawed that they must have been generated by artificial intelligence. He called the arguments relating to the meaning and interpretation of the statute tortured and tone-deaf. He labeled them superficially persuasive but said they subtly bent the law and had no merit. All of which apparently confirmed for His Honor that they had to have come from AI.

More than that, he called it lazy lawyering. He said it raised serious questions about the lawyers’ ability to practice law. He even stated he sympathized with the prosecutors for having to contend with all this. He went on to conclude that AI was undermining the ability of lawyers to represent their clients properly. 

The article reports he also threatened sanctions, including a ban on the lawyers’ ability to appear in cases in the district in the future. He held off, he said, only because he didn’t have time to hold a hearing before Thanksgiving. Busy guy.

The motion was on behalf of 70-year-old Khalwinder Singh Bhengura. Bhengura was requesting release from custody before sentencing. He was convicted last month of a felony count of false statements of citizenship to register to vote and a misdemeanor count of voting by an alien in a federal election. So this wasn’t just any motion. 

What Say The Lawyers?

Business Insider did reach out to the maligned lawyers. They admitted using AI but pointed out they followed local rules and disclosed that use to the court. They noted AI use was not prohibited. They stood by their arguments contained in their filings. 

The lawyer who signed the pleading told Business Insider that he personally wrote, reviewed, and finalized the briefs in question. And that the novel arguments came from them, not AI.

Novel Doesn’t Mean AI Slop

The case raises some troubling questions. Judges are certainly not immune to being fed up with AI slop. We all are. And it’s one thing to sanction lawyers for citing cases that don’t exist or propositions they don’t support. Deservedly so. But to disparage and threaten them for what the judge believes AI wrote? That’s another thing altogether.

Over my years practicing law, most of which were admittedly before the advent of generative AI, I can assure you I was sometimes (well, maybe more than sometimes) hit with poorly crafted briefs and arguments that made little sense, at least to me. 

Arguments that stretched precedent beyond recognition. Arguments that made me wonder whether opposing counsel had read the cases they were citing. And yes, arguments that were so poorly written I had trouble figuring out what the hell the lawyer was trying to say.

Usually, the judge’s response was not to rip the lawyers filing the briefs a new one. But to read and analyze the briefs of both sides and then make a studied decision. That’s the job of the court, not to slam a lawyer’s writing style or arguments. 

And an argument didn’t typically become frivolous merely because a judge thought it was a long shot. The fact that an argument was weak wasn’t a referendum on whether the lawyer was competent to practice law. Yes, lawyers must have a good-faith basis for their arguments and should be sanctioned if that basis isn’t there. But threatening to sanction lawyers because the judge thinks their arguments aren’t good and therefore must have come from AI worries me. 

And to top it off, he didn’t even hold a hearing to give the lawyers a chance to be heard before excoriating them. 

An AI Slop Detector?

It also worries me when a judge unilaterally concludes that a brief was written by AI or contained AI-generated arguments. That’s not a judge’s job. And who made the judge an expert at calling out AI slop? 

I read a lot of stuff that I think is AI slop and even make jokes about it on occasion. But if I were put under oath and asked whether I knew that for a fact, I would have to say I didn’t. Neither should a judge make that kind of determination and then enter it in an order. 

A Chilling Effect

There’s another reason that this is problematic. Like many of us, I use AI for brainstorming, and yes, many of the ideas it generates are downright goofy. But every now and then, it comes up with something out of the mainstream, and something I hadn’t thought of. A different approach. A new theory. Just because it may have come from AI doesn’t mean it needs to be dismissed out of hand. I certainly don’t.

But threatening sanctions has a chilling effect on lawyers’ use of AI to represent their clients competently. To try those new arguments and theories. AI use isn’t necessarily undermining their ability to represent their clients. In fact, it can be just the opposite. 

I think what this judge did was wrong. He should not have disparaged the lawyers just because he had a bad day and was angry about having to read too much AI slop. He shouldn’t have resorted to name-calling and insults. God knows we have enough of that right now. Our judges should set a better example.

I’m sorry this judge got pissed off, but that’s no excuse.

It’s Not Just About The Lawyers

Which brings me to a person who seems to have been forgotten in all this. Bhengura, the person the courts and judges are there to serve, had his liberty at stake. That’s not an academic exercise over legal writing.

That liberty may now be in jeopardy not for something he did but for something the judge thinks his lawyers did. As the case continues, the judge’s view of their conduct may color future rulings. What happens if it affects what Bhengura’s lawyers do or don’t do out of fear of retribution? Or what arguments they make? Will they worry that vigorous advocacy might be interpreted as further misconduct? 

It’s got to be in the back of their minds. And if it is, Bhengura’s the one who may ultimately pay the price.

Not to mention the court system, which may ultimately have to deal with an appeal based on the judge’s gratuitous statements.

A Simple Solution

There’s a simple way to deal with the problem: if the arguments are weak and unpersuasive, deny the motions on substantive grounds. If an argument is wrong, say why. If a case is misrepresented, identify the misrepresentation. If the filing is frivolous, apply the appropriate standards. If sanctions are warranted, provide the process the law requires.

But don’t act out of pique over how you think the lawyers used AI. That’s not judging. That’s going off the rails.


Stephen Embry is a lawyer, speaker, blogger, and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law. 

The post AI Slop Or Novel Arguments? A Texas Judge Goes Off The Rails appeared first on Above the Law.

artificial intelligence 4111582 1920
AI Slop Or Novel Arguments? A Texas Judge Goes Off The Rails 4

Hallucinations in briefs and court filings. Citations to cases that don’t exist. Citations to cases for propositions that are incorrect or that make little sense. It’s happened over 2,000 times according to Damien Charlotin’s database. And when lawyers do it, they deserve to be called out. That’s old news. 

But what happens when lawyers use AI to write a brief that reads like AI slop? Or make arguments so nonsensical that you wonder whether they came from AI rather than a human being?

Well, that’s exactly what happened recently in a federal district court case in Texas, as reported in a Business Insider article. 

The Slop Order

U.S. District Judge Mark Pittman ripped three lawyers to shreds not for hallucinated cases but for filing motions containing what he described as blatant AI slop, according to the article. 

The judge concluded the arguments were so weak and legally flawed that they must have been generated by artificial intelligence. He called the arguments relating to the meaning and interpretation of the statute tortured and tone-deaf. He labeled them superficially persuasive but said they subtly bent the law and had no merit. All of which apparently confirmed for His Honor that they had to have come from AI.

More than that, he called it lazy lawyering. He said it raised serious questions about the lawyers’ ability to practice law. He even stated he sympathized with the prosecutors for having to contend with all this. He went on to conclude that AI was undermining the ability of lawyers to represent their clients properly. 

The article reports he also threatened sanctions, including a ban on the lawyers’ ability to appear in cases in the district in the future. He held off, he said, only because he didn’t have time to hold a hearing before Thanksgiving. Busy guy.

The motion was on behalf of 70-year-old Khalwinder Singh Bhengura. Bhengura was requesting release from custody before sentencing. He was convicted last month of a felony count of false statements of citizenship to register to vote and a misdemeanor count of voting by an alien in a federal election. So this wasn’t just any motion. 

What Say The Lawyers?

Business Insider did reach out to the maligned lawyers. They admitted using AI but pointed out they followed local rules and disclosed that use to the court. They noted AI use was not prohibited. They stood by their arguments contained in their filings. 

The lawyer who signed the pleading told Business Insider that he personally wrote, reviewed, and finalized the briefs in question. And that the novel arguments came from them, not AI.

Novel Doesn’t Mean AI Slop

The case raises some troubling questions. Judges are certainly not immune to being fed up with AI slop. We all are. And it’s one thing to sanction lawyers for citing cases that don’t exist or propositions they don’t support. Deservedly so. But to disparage and threaten them for what the judge believes AI wrote? That’s another thing altogether.

Over my years practicing law, most of which were admittedly before the advent of generative AI, I can assure you I was sometimes (well, maybe more than sometimes) hit with poorly crafted briefs and arguments that made little sense, at least to me. 

Arguments that stretched precedent beyond recognition. Arguments that made me wonder whether opposing counsel had read the cases they were citing. And yes, arguments that were so poorly written I had trouble figuring out what the hell the lawyer was trying to say.

Usually, the judge’s response was not to rip the lawyers filing the briefs a new one. But to read and analyze the briefs of both sides and then make a studied decision. That’s the job of the court, not to slam a lawyer’s writing style or arguments. 

And an argument didn’t typically become frivolous merely because a judge thought it was a long shot. The fact that an argument was weak wasn’t a referendum on whether the lawyer was competent to practice law. Yes, lawyers must have a good-faith basis for their arguments and should be sanctioned if that basis isn’t there. But threatening to sanction lawyers because the judge thinks their arguments aren’t good and therefore must have come from AI worries me. 

And to top it off, he didn’t even hold a hearing to give the lawyers a chance to be heard before excoriating them. 

An AI Slop Detector?

It also worries me when a judge unilaterally concludes that a brief was written by AI or contained AI-generated arguments. That’s not a judge’s job. And who made the judge an expert at calling out AI slop? 

I read a lot of stuff that I think is AI slop and even make jokes about it on occasion. But if I were put under oath and asked whether I knew that for a fact, I would have to say I didn’t. Neither should a judge make that kind of determination and then enter it in an order. 

A Chilling Effect

There’s another reason that this is problematic. Like many of us, I use AI for brainstorming, and yes, many of the ideas it generates are downright goofy. But every now and then, it comes up with something out of the mainstream, and something I hadn’t thought of. A different approach. A new theory. Just because it may have come from AI doesn’t mean it needs to be dismissed out of hand. I certainly don’t.

But threatening sanctions has a chilling effect on lawyers’ use of AI to represent their clients competently. To try those new arguments and theories. AI use isn’t necessarily undermining their ability to represent their clients. In fact, it can be just the opposite. 

I think what this judge did was wrong. He should not have disparaged the lawyers just because he had a bad day and was angry about having to read too much AI slop. He shouldn’t have resorted to name-calling and insults. God knows we have enough of that right now. Our judges should set a better example.

I’m sorry this judge got pissed off, but that’s no excuse.

It’s Not Just About The Lawyers

Which brings me to a person who seems to have been forgotten in all this. Bhengura, the person the courts and judges are there to serve, had his liberty at stake. That’s not an academic exercise over legal writing.

That liberty may now be in jeopardy not for something he did but for something the judge thinks his lawyers did. As the case continues, the judge’s view of their conduct may color future rulings. What happens if it affects what Bhengura’s lawyers do or don’t do out of fear of retribution? Or what arguments they make? Will they worry that vigorous advocacy might be interpreted as further misconduct? 

It’s got to be in the back of their minds. And if it is, Bhengura’s the one who may ultimately pay the price.

Not to mention the court system, which may ultimately have to deal with an appeal based on the judge’s gratuitous statements.

A Simple Solution

There’s a simple way to deal with the problem: if the arguments are weak and unpersuasive, deny the motions on substantive grounds. If an argument is wrong, say why. If a case is misrepresented, identify the misrepresentation. If the filing is frivolous, apply the appropriate standards. If sanctions are warranted, provide the process the law requires.

But don’t act out of pique over how you think the lawyers used AI. That’s not judging. That’s going off the rails.


Stephen Embry is a lawyer, speaker, blogger, and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.