Courts on AI / Fabricated or misquoted citations
Wilkinson v. Schmelzer
Court of Appeals of Iowa · State appellate courts · IA · · No. 25-0799 · opinion
The Court of Appeals of Iowa affirms a custody order and addresses a sanctions request based on alleged AI use. The court finds no hallucinated case, says it cannot determine whether a nonexistent appellate rule citation is a typo or unchecked AI output, and declines AI-use sanctions.
- Disposition
- The court affirmed, declined AI-use sanctions, and awarded appellate attorney fees and costs on other grounds.
- Recorded conduct
- Self-represented appellant cited nonexistent Iowa R. App. P. 6.1004(3), but the court could not tell whether it was a typo or unchecked AI output. — outcome: other; actor: litigant in person
- Canonical record
- Incident tracker entry — verified outcome, regulator disposition and related rules
What the court wrote
Verbatim from the court's document (public domain). Ellipses mark omitted text; nothing is paraphrased.
In her resistance to Brett’s motion for remand and request for sanctions, Lacey asked our supreme court to impose sanctions on Brett for purported use of generative AI in his limited-remand motion and briefs. ... The twist is that, absent the presence of a hallucinated case, it is difficult to differentiate poor human drafting from unverified generative AI. Here, we can find no hallucinated case. And it is impossible for us to say with certainty whether Brett’s single reference to the nonexistent “Iowa R. App. P. 6.1004(3)” was a simple typo or the result of unchecked AI use. Thus, we decline to apply sanctions for Brett’s purported use of AI. But regardless of whether Brett’s briefs were directly drafted by him or a faraway datacenter, he is bound by the same duty to verify his briefs’ accuracy and compliance with our appellate rules.
Authorities cited in the passage: Luke v. Department of Health & Human Services, 2025 WL 2237311 (Iowa Ct. App. Aug. 6, 2025) · Iowa R. App. P. 6.1004(3) · Iowa R. App. P. 6.903
compiled by the Cognesio LLP editorial teammethodas of 2026-09-07v0.1.1coding CC BY 4.0JSON Hugging Face GitHub
Hugging Face