October 18, 2026 (in 16 days): Tenth Circuit: comments close on 2027 rules, incl. proposed Rule 46.5(B)(4) generative-AI review →

Barari v. Burgess

Arizona Court of Appeals, Division One · Ariz. Ct. App. · Arizona bar guidance

Pro-se party

Conduct

Self-represented mother filed a brief in which 24 of 52 citations were deficient, one to a nonexistent case, then gave shifting accounts of her AI use.

Consequence

Deficient citations struck as an ARCAP 25 sanction, leave to amend denied, arguments waived, and the order affirmed.

Lesson

She admitted only minimal AI use; weighing the pattern of citation errors, the court found her insistence that she did not use it for legal research not credible.

Court sanction

Verified September 28, 2026

Citation
Barari v. Burgess, No. 1 CA-CV 25-0815 FC (Ariz. App. June 26, 2026) (mem. decision)
Decided
June 26, 2026

Summary

Self-represented appellant Nastaran Barari ("Mother") appealed a Maricopa County order denying her request to modify a 2025 dissolution decree. Her opening brief cited nothing in the record, and "[o]f the fifty-two citations in her brief, twenty-four are deficient," including cases cited for principles they do not support, misstated rules, misrepresented facts and holdings, and one case that "simply does not exist, commonly known as a 'hallucinated' case." The court ordered her to appear at a show-cause hearing and testify under oath, and allowed a written explanation. She insisted she wrote the brief and did all research herself on Westlaw at a law library, "only to later admit that she did use generative artificial intelligence, but only minimally." She attributed the nonexistent case to an inadvertent mis-citation of another case, which the court found was itself unreported, used different names, and did not support her proposition. The court found her answers "evasive, evolving, and inconsistent" and her insistence that she did not use generative AI for legal research "not credible."

AI tool:
Generative AI (the court found not credible the appellant's insistence that she did not use it; she later admitted using it "only minimally"; tool not identified)
Sanction amount:
Non-monetary: motion to amend the opening brief denied and, as a sanction under ARCAP 25, each inaccurate and false legal citation struck; arguments deemed waived
This case summary is informational only. Verify the underlying opinion or order against the primary source before relying on it in any filing or client matter.

What sanction did the court impose?

In a memorandum decision (Furuya, J., joined by Jacobs, P.J., and Morse, J.), the court held that Mother's "submission of false citations warrants sanctions under ARCAP 25." It denied her motion, made at the show-cause hearing, to amend her opening brief and, as a sanction, struck "each inaccurate and false legal citation at issue." It also disregarded all facts stated without record citations. With the citations struck, the brief lacked supporting authority, the court held that Mother had not made a "bona fide and reasonably intelligent effort" to comply with ARCAP 13, deemed her arguments waived, and affirmed. To the extent she sought review of the March 2025 decree itself, the court held it lacked jurisdiction because her September 2025 notice of appeal was untimely. No monetary sanction or fee award was imposed; the appellee was also self-represented.

Why does Barari v. Burgess matter for law firms using AI?

The decision is notable for how the court reached its AI finding. Mother did not concede AI use at first; she testified that she researched on Westlaw at a law library and checked every citation, and explained the nonexistent case as a mix-up with another case. The court tested that account against the errors themselves. It observed that “combination of case names with unrelated reporters and pin citations is commonly encountered in legal materials drafted using generative artificial intelligence,” and that “[m]istaking an unreported decision for a differently named and reported opinion is not a common or likely error” for a careful researcher. It therefore found her denial “not credible.”

The sanction was procedural rather than monetary. Relying on Ramos v. Nichols, the court treated ARCAP 25 as authority for sanctions fitted to the circumstances of the case and meant to discourage similar conduct, and struck each inaccurate and false citation, denying Mother’s request to amend her brief at the show-cause hearing. The court described the citation errors as “misrepresentations of the law to the court, which we will not tolerate.”

For practitioners, the case uses steps that also appear in Division One’s later published Acciavatti opinion: a show-cause hearing at which the filer testified under oath, and a remedy that strikes the inaccurate and false citations from the brief. Barari adds a credibility finding against the filer’s account of her research. Firms documenting compliance may wish to consider that a court may infer AI use from the character of citation errors even where the filer denies it.

Implications for your firm

Operational steps a firm reading this case may wish to consider documenting. Strategic and rule-application calls belong to your firm's attorneys.

  • Document the research path for each authority in a filing. The court tested the appellant's claim of Westlaw research against the error pattern and found the claim not credible.
  • Consider that striking citations can end an appeal without any fine: once the false citations were removed, the brief lacked support and the arguments were waived.
  • Consider that filers whose citations are questioned may wish to prepare for the possibility of a show-cause hearing with sworn testimony; here the court questioned the filer under oath and weighed her credibility before choosing the sanction.

Sources

Primary sources