Background
Gloria Dolores B. Acciavatti died in June 2023 at age 93, leaving three adult children: Kim Dineen, Jeffrey Kotchka, and Jill Shibata. Upon reviewing estate records, Dineen discovered that all assets were titled in Kotchka’s name. In February 2024, Dineen petitioned for formal probate and appointment as personal representative of the estate. Kotchka and Shibata opposed the petition. After a hearing in July 2025, the trial court appointed Dineen as personal representative and issued an order for formal probate of the will.
Kotchka timely appealed the trial court’s decision, raising arguments regarding due process, fair hearing rights, and alleged material errors of fact and law. However, his opening brief contained multiple deficient legal citations. Upon review, the court of appeals discovered that Kotchka had cited eight cases, six of which were deficient—including misstatements of legal principles, incorrect page citations, and misrepresentations of case facts. Most significantly, two of the cited cases simply did not exist.
When ordered to appear at a show cause hearing and explain the citations, Kotchka admitted that he had relied on “artificial-intelligence-assisted research tools” for his brief and had not independently verified the citations using legal research tools, public law libraries, or professional legal assistance.
The Court’s Holding
The court affirmed the trial court’s order on the ground that Kotchka’s opening brief violated Arizona Rule of Civil Appellate Procedure 13 by failing to provide accurate legal citations. Because the brief contained no other supportive legal authority for its arguments, the court struck the deficient citations and treated this as a waiver of all issues raised on appeal. The court held that Kotchka failed to make “a bona fide and reasonably intelligent effort” to comply with appellate rules, which require all litigants—whether represented by counsel or self-represented—to adhere to the same standards.
The court determined that Kotchka’s use of hallucinated case citations constitutes sanctionable abuse of Gen-AI tools. The court emphasized that using Gen-AI is not inherently objectionable and can assist self-represented litigants, provided users verify all citations and facts. However, when users fail to check whether cases are real or accurately cited, they abuse Gen-AI tools and misrepresent the law to the court. The court found that Kotchka’s submission of false authorities—whether intentionally or unintentionally—violates A.R.S. Section 12-349, ARCAP 25, and the court’s inherent powers to manage its affairs. Crucially, the court held that lack of legal training and lack of intent to deceive do not excuse such misrepresentation.
The court imposed sanctions against Kotchka, awarding Dineen her reasonable attorneys’ fees and costs on appeal under A.R.S. Section 12-349 and ARCAP 25, specifically for the resources she expended in identifying and correcting the hallucinated and misrepresented citations in Kotchka’s brief.
Key Takeaways
- Gen-AI tools can lawfully assist in legal work only when used to perform tasks like editing, grammar checking, and document summarization, provided all citations and facts are independently verified before filing.
- Citing hallucinated cases—cases that do not exist—or misrepresenting existing case law through Gen-AI-generated material constitutes abuse of Gen-AI and is sanctionable conduct in Arizona courts.
- Self-represented litigants are held to identical standards as licensed attorneys regarding accuracy and compliance with procedural rules; lack of legal training is no defense to filing false legal authorities.
- Gen-AI’s tendency to “hallucinate” nonexistent cases arises from the training and structure of AI programs and cannot be cured by users relying on confidence in the tool; reasonable diligence requires independent verification using trusted legal research sources.
- Misrepresenting the law to the court through hallucinated citations constitutes bad faith conduct regardless of intent, and the absence of intent to deceive does not absolve litigants from sanctions.
Why It Matters
This decision represents the Arizona Court of Appeals’ first published ruling directly addressing Gen-AI hallucinations in appellate briefs, establishing clear standards at a time when both attorneys and self-represented litigants increasingly use AI tools to draft legal documents. The court made clear that Gen-AI’s known tendency to fabricate nonexistent legal authorities cannot excuse litigants from their duty to verify citations before filing—and that courts will sanction those who fail to do so. This ruling directly confronts a rapidly emerging problem: courts nationwide have reported hundreds of cases in which litigants have cited fake cases generated by AI, creating a systemic threat to the integrity of the legal process.
The decision also reaffirms that self-represented litigants must comply with the same rules of procedural and substantive law as licensed attorneys, while acknowledging that such litigants face practical challenges in accessing legal research tools. By imposing sanctions tied specifically to the extra work required to identify and correct hallucinated citations, the court signals that misuse of Gen-AI imposes real costs on opposing parties and the judicial system. The ruling serves as a warning to both pro se litigants and attorneys that blind reliance on Gen-AI output—without verification against authoritative legal databases or trusted sources—will not be tolerated and may result in sanctions including attorneys’ fees.