A lot has been written about the dangers and problems of Artificial Intelligence (AI) in the courts. First it was lawyers who let AI write their appellate briefs, without noticing the AI’s hallucinations. Some lawyers even had bar inquiries arising from their failure to double-check their AI. And lawyers blaming their staff for failure to double-check didn’t improve their success rate with angry judges, either.
Then it was opposing lawyers who couldn’t collect fees or sanctions because they hadn’t noticed the hallucinations, either. Most recently, a judge ordered sanctions against the opposing attorneys — the ones who didn’t use AI or include its hallucinations in their pleadings. The court’s reasoning: they also had a duty to read their opponents’ work and discover — and disclose — the problems.
Is AI in the courts affecting elder law, too?
Not as much. But it’s coming.
Why have elder law attorneys not seen as much trouble with AI? It’s mostly because we don’t usually spend much time writing briefs filled with legal citations. The practice of elder law is far less likely to lead us to courts at all. And appellate proceedings involving “our” issues are even rarer.
That’s not to say we don’t have to pay attention to AI hallucinations. We do. But if an elder law attorney asks for AI’s help in figuring out a statute or a rule, we’re far more likely to be able to spot the mistakes. We’re also far more likely to know the basic rules, or even the specific answer, but needing to find the citation to that authority.
AI use in the elder law context just doesn’t provide the same opportunity for hallucinations, the most common complaint against the practice in popular literature. But that doesn’t mean it doesn’t happen. In fact, a recent Arizona Court of Appeals case provides a good illustration of what can happen when AI in the courts turns bad.
Gloria Acciavatti’s estate
Gloria Acciavatti died in 2023, at the age of 93. She left three children: Kim Dineen, Jeffrey Kotchka, and Jill Shibata. Ms. Dineen initiated probate proceedings, only to discover that all of her mother’s assets had been titled in Mr. Kotchka’s name. She asked the probate court to give her authority to pursue her brother for the lifetime transfers that put his name on all of the assets.
Both of her siblings opposed her, but the probate court appointed her as personal representative so that she could continue her exploration. Mr. Kotchka appealed that ruling.
Mr. Kotchka did not have a lawyer in the appeal process. But he managed to put together his opening brief to file with the appellate court. Problem was, he had included eight case citations, and six of them were “deficient”, according to the Court of Appeals. Among the problems, according to the court: his citations “include the use of cases for principles that are not supported by the case, mis-citation to wrong pages or paragraphs, misstatement of rules enunciated in cases, and misrepresentation of case facts.” “Most concerning,” wrote the court, “the brief cites two cases that simply do not exist.”
The appellate court ordered Mr. Kotchka to explain himself. He responded that he had used AI to write his brief and he blamed it for the errors. When he appeared before the judges, he testified that he did not consult attorneys, or use regular law libraries or publicly-available resources. He had not even reviewed the court’s own suggestions for self-represented appellants.
The upshot
In their review of AI in the courts, the appellate judges noted that AI itself is not the problem. In fact, competent lawyers and laypersons might use AI to help them locate authorities, check grammar, form clearer arguments and help in other ways. But the litigant — just like the lawyer they might hire to do the legal work — must check their work.
Mr. Kotchka was assessed a penalty for his misuse of AI in the courts. He was ordered to pay the a portion of his sister’s legal fees. And that, ruled the court, was true even if he did not intentionally mislead the judges or understand what he had done wrong. Estate of Acciavatti, Dineen/Shibata v. Kotchka, July 15, 2026
So what does that tell us about elder law and AI?
Not much. It’s a relatively rare event to see an opportunity for hallucinations to appear in court proceedings in the elder law world. But it can happen, has happened and will happen again.
But even if it is rare, it is instructive. AI is not itself an evil influence (though we appreciate that reasonable minds might differ on this score). It is the abuse and misuse of AI — usually occasioned by the ease of delegating to the appearance of ominiscience — that is problematic. AI can be and is often a useful tool in the practice of law generally, and elder law specifically. But we’re a long way from being able to simply turn the arguments over to the artificial lawyer.
One Response
I have practiced law since 1981. While the dangers and risks of AI are real, it is also an example of “the more things change, the more they stay the same.” For example, our research and writing professor in law school warned us not to rely on the annotations or other summaries of cases found in digests, but to read the case itself, as many times, the headnote does not reflect the actual holding of the case.
I have found the best use of AI occurs when I do the research and prepare a draft myself, using traditional methods, and then run the draft through AI to refine it into a final version. Even then, I carefully examine the AI product. Never ask AI a legal question and rely without investigation on the answer it provides. I did that once, and it cited an Arizona statute that was directly on point and answered the question precisely. The only problem was that there was no such Arizona statute. AI had made up the statute number out of thin air. It frequently cites non-existent cases. On a non-legal point, one version of AI thinks Francis is still pope and that there has not been a pope named Leo in over 100 years. That raises a question in my mind in how up-to-date AI might be.