Submit incident
Documented

Fabricated AI Citations in a Hoverboard Lawsuit Left Injured Plaintiffs at Risk

January 1, 2025
Curated by Team Raidu · Reviewed by Shiva Ganesh
aiaaic:AIAAIC1919View source ↗
LinkedInX

What happened

In February 2025, attorneys representing plaintiffs in a product liability case against Walmart were found to have submitted court filings that contained fabricated legal citations generated by an AI system. The underlying lawsuit, originally filed in June 2023, centered on a house fire the plaintiffs alleged was caused by a defective hoverboard. Before the case could be heard on its merits, the court had to address something that had nothing to do with hoverboards or Walmart.

The false citations appeared in a motion in limine, a pretrial filing used to exclude specific evidence or arguments from trial. Attorneys working on the motion had used ChatGPT to help draft it. The AI system returned case citations that looked authoritative: names, apparent legal reasoning, the surface texture of real precedent. None of the cases existed. The lawyers acknowledged afterward that they had relied on the output without verifying any of the references against an actual legal database before submitting the motion.

The disclosed facts made a bad situation worse. Two of the attorneys involved had not read the motions before they were filed. The fabricated citations cleared the entire legal team without anyone checking whether the cases were real. The pattern is consistent with similar incidents in other courts: a generative tool produces confident-looking research, no one verifies it against an authoritative source, and the error surfaces only when a judge or opposing counsel searches for the cited case and finds nothing.

The people who bear the immediate cost of that failure are the plaintiffs. People who filed suit over injuries and property damage from a house fire now face the possibility that their case is compromised or delayed because of procedural errors introduced at the drafting stage. The attorneys who submitted the unverified citations saved themselves the time it would have taken to check a handful of references while exposing their clients to that risk.

Nothing in the current filing process creates a record of how a legal document was produced, or what a human confirmed before it was submitted. A court receives a brief and has no way to know whether its citations came from a trained attorney working from a verified database or from a language model generating plausible-looking text with no connection to the legal record. A provable record of what a system did and what a human verified before submission would have caught the fabricated references before they reached a judge, and would give courts a consistent basis for evaluating how AI tools are being used in proceedings before them.

Reported impact

Affected parties
Not publicly disclosed
Harm type
Not publicly disclosed
Scale
Not publicly disclosed
Financial impact
Not publicly disclosed
Regulatory action
Not publicly disclosed

Classification

Organization
Not publicly disclosed
AI system
Not publicly disclosed
Industry
Not publicly disclosed
Country
Not publicly disclosed
Provider
Not publicly disclosed
Incident type
Not publicly disclosed

Relevant governance controls

Governance control mapping is not available for this record.

  • No controls mappedNot publicly disclosed

Control mapping is analytical. It does not state that any control would have prevented the incident.

Sources and evidence

This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.

AIAAIC Repository
Also catalogued in
Fabricated AI Citations in a Hoverboard Lawsuit Left Injured Plaintiffs at Risk
2025