Background
Nastaran Barari and Kemmy D. Burgess married in December 2014 and had two children together. In 2023, after nine years of marriage, the spouses separated. Mother filed a Petition for Dissolution of Marriage, and the parties heavily litigated Father’s financial disclosures over two years. In March 2025, the superior court issued a Decree for Dissolution resolving parenting time, child support, legal decision-making, and requests for sanctions and attorneys’ fees.
Mother did not appeal the Decree itself but subsequently filed various post-judgment motions, including motions to compel financial disclosures, to modify the Decree, and to enforce parenting time—all of which were denied. Mother appealed the order denying her request to modify the Decree.
The Court’s Holding
The Arizona Court of Appeals affirmed the trial court’s decisions. The court found that Mother’s appellate brief was fundamentally deficient under Arizona Rule of Civil Appellate Procedure 13. Of fifty-two citations in her brief, twenty-four were deficient or false, including one citation to a case that does not exist—commonly known as a “hallucinated” case.
The court concluded these citation errors were consistent with abuse of generative artificial intelligence without proper review. When ordered to appear at a show-cause hearing and testify under oath, Mother initially denied using AI, then admitted to “minimal” use, and provided implausible explanations for the errors. The court found her testimony not credible, particularly given that the pattern of errors (mixing unrelated reporters with similar case names) is common in AI-generated legal materials.
As a sanction under ARCAP 25, the court struck all false citations from the brief and deemed Mother’s arguments waived for failure to comply with appellate procedure rules. The court emphasized that it would not tolerate misrepresentations of law to the court or develop arguments for parties who fail to comply with procedural requirements.
Key Takeaways
- Appellants must comply with appellate procedure rules requiring accurate citations to legal authority and the record; failure to do so constitutes waiver of arguments.
- Courts will impose sanctions for briefs containing numerous false citations, particularly when evidence suggests generative AI was used without adequate verification.
- Attorneys using AI for legal research must thoroughly review and verify all citations before filing; the court will not credit implausible explanations for AI-generated errors.
- False citations constitute misrepresentations of law to the court and undermine the judicial process.
Why It Matters
This decision addresses the emerging problem of generative AI in legal practice. As attorneys increasingly use AI tools for legal research and writing, courts are establishing clear consequences for failure to verify AI-generated content. The court’s willingness to strike arguments and impose sanctions for false citations demonstrates that professional responsibility standards apply to AI-assisted legal work—users cannot delegate accuracy to the technology.
The decision signals that courts will scrutinize patterns of citation errors consistent with AI generation and will hold attorneys accountable when they misrepresent their due diligence or fail to adequately review AI outputs before filing. This case serves as a cautionary example of the risks of relying on generative AI without proper human verification and review in appellate practice.