To be clear, the judge probably did not let AI decide this case. The plaintiff sued Nevada family court judge Mari Parladé, accusing her of farming out her opinion to AI based on the judge reading from a prepared script, stumbling over a word while reading it, and accidentally uttering “appearance of any impartiality” when reading “appearance of impropriety.”
That’s not particularly robust gruel.
But in any event, Judge Gloria Navarro of the District of Nevada dismissed the case, determining that the federal court doesn’t even need to reach the substance of Judge Parladé’s alleged AI use because judicial immunity would apply even if the judge relied “wholly on artificial intelligence” to issue a ruling. A judicial act is a judicial act.
Professor Eugene Volokh flagged the decision, noting that while even complete AI reliance can become an issue on appeal or give rise to disciplinary proceedings, this case stands for the proposition that it cannot escape the fact that choosing to use AI — even if that happened here — would be a judge’s decision and therefore shielded from federal liability.
The four-factor test asks if the act is a normal judicial function, if the events occurred in the judge’s chambers, if the controversy centered around the case pending before the judge, and if the events at issue arose out of confrontation with the judge in his or her official capacity. Nothing about that test precludes a judge from deciding that it’s reasonable to schedule a Claude cron job to clear the docket. It would be ill-advised, just as it’s professionally suspect to hand over the entire civil docket to clerks, but it’s still a choice within the scope of the judge’s… well, judgment.
When two federal judges issued orders with fabricated quotes, it took a Senate Judiciary Committee inquiry to shake loose the admission that an intern used ChatGPT and a clerk used Perplexity. That triggered professional embarrassment, but it did not open the judges up to a collateral lawsuit.
But, again, there’s not much to suggest the judge in this case actually used AI — much less delegated the whole case to the wisdom of the bots. From the complaint:
Notably, on Aug. 15, 2023, the judge read aloud from a prepared script—which tends to suggest A.I. use. More notable still, while reading aloud, the judge stumbled over the words—which suggests the judge is not the author of the prose read aloud in court that day, which again, tends to suggest A.I. use.
Judges have read from prepared remarks since before the invention of movable type. And they’ve botched those dramatic readings for just as long.
The complaint characterizes the misspoken “appearance of any impartiality” as “a classic, Freudian slip-of-the-tongue, A.I. style.” But since AI is not saddled with an Oedipal complex — as far as we can tell — anything “Freudian” would tend to point to a human author. AI can hallucinate, but it would not screw up impropriety and impartiality in that context.
There are more allegations and they’re all equally dubious.
This is not the first time the plaintiff — a California-licensed attorney appearing pro se here — brought a federal suit against the people running his state family court case. According to Magistrate Judge Nancy Koppe’s report, it’s at least his tenth challenge, with a record of “success in none of them.”
In any event, of all the reasons not to hand over judicial rulings to AI, fear of a lawsuit isn’t one of them.
P.S. I still think Justice Roboto was robbed in our Law Revue contest.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news.
The post Judicial Immunity Still Applies Even If The Judge Let AI Decide The Case appeared first on Above the Law.