I've had artificial intelligence on the brain (get it?) this week, after seeing a recent high profile incident involving the lawyers for Mike Lindell, founder of My Pillow.

In 2022, Mr. Lindell, My Pillow, and another company were sued for defamation in Colorado by a former executive of Dominion Voting Systems after Mr. Lindell allegedly claimed falsely that the executive had admitted to rigging the 2020 election.

(You remember the 2020 election, right? Good. Because this post is not about that.)

The defamation lawsuit against Mr. Lindell and his co-defendants is scheduled to go to trial one month from today.

In February, the plaintiff's attorneys filed a motion with the court. As one would expect, Mr. Lindell's attorneys opposed the motion. But the defense brief was full of AI "hallucinations." Either made-up case law, or actual court decisions that didn't say what the brief said they said. The plaintiff's attorneys apparently didn't notice because their reply didn't mention it.

On the other hand, the judge (Nina Wang, a Biden appointee) did notice. At a pretrial conference last week, the judge started asking Mr. Lindell's lead counsel (we'll call him "The Partner") about the hallucinated case law in his brief. The Partner didn't respond very well to the judge's questions. The Partner admits this. Then the judge flat-out asked him whether the brief was written with generative AI, and he said yes. Then the judge got mad and issued an order directing The Partner to show cause as to why she shouldn't sanction him, his associate (we'll call her "The Associate"), and the defendants, and refer the attorneys for disciplinary proceedings.

This week, The Partner and The Associate filed a pleading and sworn statements and other exhibits telling their story, and apologizing to the court and to opposing counsel. (The Partner's apology struck me as a bit hostile, but he did apologize. The Associate seemed to be genuinely sorry.)

Here's what the attorneys said about the AI. The brief was due while The Partner was on vacation in Mexico, but he was in touch with The Associate, who was back home and doing most of the work. According to both attorneys, The Partner had prepared an outline using AI, but while he was in Mexico, he and The Associate went through the draft together and made significant revisions that included correcting the AI errors.

Then, according to both attorneys, when it was time to file, The Associate accidentally filed . . . the draft. Instead of the final version that was AI-hallucination-free.

(Dear Readers, this does sometimes happen. I had a case in pre-AI days where my opponent accidentally filed a draft of his brief instead of the final version. In my case, though, he realized it immediately and asked to be allowed to replace the "draft" brief with the "final" version. Our side had a chuckle over it, but it was not a big deal.)

Then, according The Associate and The Partner, neither one of them ever looked at the filed brief again, so in the almost two months before the pretrial conference, they believed that the final version was the one that had been filed with the court. The Partner didn't even bring a copy of the brief to the pretrial conference. Which was why the judge's questions about the brief caught The Partner off guard and he couldn't respond appropriately. Because up until that moment, he thought the final version of the brief was the version that the judge had. 

The judge is not impressed. She now wants the originals (with metadata) of the emails exchanged between The Partner and The Associate before the brief was filed, and all versions of the brief. The deadline is this Monday, May 5.

The attorneys for the plaintiff said it was kinda hard to believe that the attorneys for Mr. Lindell never so much as looked at the brief that they filed before going into the pretrial conference. That's a fair point.

AI is fallible!

This is yet another true story proving that AI is a blessing but also a curse. It can be a great resource for many things, like helping to compose a thank-you note to your Aunt Millie for the argyle socks she gave you for Christmas.

By the way, in case you are curious -- and I know you are! -- here is the thank-you note that Grok composed for me in about 2 seconds last night:

Dear Aunt Millie,

I hope this note finds you well! I wanted to take a moment to thank you for the wonderful argyle socks you gave me for Christmas. They are not only stylish but also incredibly cozy! I truly appreciate your thoughtfulness and generosity.

Every time I wear them, I'll think of you and the lovely memories we've shared during the holidays. I'm looking forward to our next get-together!

Sending you lots of love and warm wishes.

Thank you again!

With love,

[Your Name]

Pretty good, huh? Better than the thank-you notes I compose myself.

But if you're doing something more important than writing a thank-you note -- like making a medical diagnosis, writing your doctoral dissertation, or submitting a brief on behalf of your pillow magnate client whose defamation trial is imminent -- you need to check behind the AI. Lawyers should know this, but because they get burned for it all the time, apparently they don't. The courts are really cracking down on lawyers who submit pleadings with AI-hallucinated legal authority.

I'm not saying that Mr. Lindell's attorneys did not check and correct the AI, as they claimed. If the emails and prior versions of the brief show that The Associate really did accidentally file a rough draft of the brief (and that a good, corrected version really was done and ready to go when the bad version was accidentally filed), then the attorneys' apologies, IMO, should be accepted. The material they presented to the court with their response to the judge's Show Cause Order seems to indicate that this is what happened. 

On the other hand, if the metadata doesn't back them up . . . ugh. The attorneys for Mr. Lindell -- not to mention their clients -- could be in for a bad trip.

I'll keep you posted.