Skip to main content

My worst nightmare as an AI using lawyer - 23 page opinion out of Mississippi

  • June 19, 2026
  • 0 replies
  • 7 views

EllenSemble

This is INSANE. Has anyone seen this? Any Mississippi lawyers around to comment? I just read it this morning and was blown away.

If you haven't heard of it, here’s the article. And here’s the opinion. Basically there is a case in MS federal court and both parties had hired out of state lawyers (I understand why MS wants that emphasized but I think it’s a silly distinction to make here given the actual story) and of course had local counsel. Out of state counsel apparently wrote the briefs, since it was likely “their” case, and the local counsel for both sides approved and signed the brief. 

Full disclosure, I have never had to do federal litigation in a federal court in a different state that I had to PHV into, since my practice is pretty local and I keep it that way. But I know that most federal districts require you to have local counsel, and they have to sign off on the briefs and many states require them to appear in person as well for at least some of the appearances. 

Here, out of state counsel wrote the brief and response brief, and local counsel on each side attested that they reviewed it. Clearly not, because both briefs had AI hallucinations and the judge was mad. She used Rule 11 to chastise the lawyers for their failures to properly verify and then disqualified all 4 out of state counsel, leaving it to the local lawyers. During the show cause hearing, one of the lawyers said she didn’t know that AI could hallucinate.

I was blown away by that … until I kept reading the opinion. That same lawyer had been cited in other jurisdictions for AI miscites!! I can’t. How does this even happen? Why lie to a judge like that? I’ve been practicing a long time and I still find myself incredulously asking that. I don’t understand risking your livelihood like that. 

All of that is wild. Imagine being the lawyer who agreed to pair up with a colleague who assured you that it was “their” case and they’d do all the work, and they did, but they did such a poor job that now they’re off the case and it’s “yours” and you have to proceed. Imagine knowing your colleague will lie/obfuscate before the court about not knowing AI could hallucinate when they’ve gotten in trouble for that before. I just can’t with any of it.

I use AI. I use legal AI. I use AI in my cases. I tell my clients I’m using AI. I have no problem doing any of that, because I know how I’m using AI and that I’m using it carefully and ethically. I even wrote about that here in this lounge before. When I used Archie in my Smokeball, I’m comfortable because I know it’s only working on the files I put in there so it won’t hallucinate AND because I check its work myself!! That’s why I’m the lawyer and my name is on the door! I don’t understand blindly turning a brief over to an AI, especially an open environment AI that trains on anything, when you wouldn’t trust a brand new associate to write a brief and submit it without your review. Plus, it is so easy to verify that Archie is doing its job correctly, often as simple as doing a Ctrl+F or just reviewing the source document. As for when I use Claude, I still treat it as if it’s a bunch of lawyers and also some laypeople who worked together to create the output, so I definitely have to verify it. I just don't understand being so irresponsible with AI use, especially when we have a duty to our clients and a duty to follow our ethics and conduct rules. Oh, and to NOT LIE TO A JUDGE. 

If you want to read the opinion, I recommend it. I’m linking it here. It’s 23 pages, so strap in, but it’s worth it. The judge used Rule 11, predictably, and really laid it out well. Phew. I still wake up in cold sweats sometimes imagining that I’m on the receiving end of a Rule 11 issue. Yikes yikes yikes.