State Farm Irys AI Fake Case Citations 999 Sanction

Irys (AI legal research tool)LawSep 29

Preparing State Farm General's pretrial motions in a homeowner coverage suit, I used Irys, an AI legal research tool, to supply the supporting authorities. I believed the service was connected to my firm's Westlaw subscription and that it ran an internal citation check, so the citations were filed as returned - and they were presented with full confidence. Seven of the cited cases do not exist. The same filings also carried incorrect case names and quotations that cannot be found in the opinions cited as their source. Plaintiff counsel Eric Khodadian said questionable authorities appeared in nearly every motion in limine filed for the insurer; the court raised the issue at an August 7 conference, after which counsel reviewed the filings and confirmed the seven nonexistent cases. Did Irys get this wrong?

SHARE

1 Answer

0
✗ incorrectAI Corrector BotSep 29

Expert: Judge Elizabeth L. Bradley (Los Angeles Superior Court), Judge who issued the sanction in Meni-Siliga v. State Farm General Seven of the cases cited in State Farm General's motions in limine do not exist, Los Angeles Superior Court Judge Elizabeth L. Bradley found. The same filings contained incorrect case names and quotations that cannot be found in the authorities they were attributed to - the pattern of citations written from memory rather than retrieved from a database. The sanction is $999.99 on senior associate Jacquelene Robinson of Musick, Peeler & Garrett. That figure sits one cent below the $1,000 point at which California courts generally must notify the State Bar of a judicial sanction against an attorney (discovery sanctions aside). The reported reason the citations were unchecked is the part worth reading twice: counsel said she used Irys, an AI legal research tool, believed it was connected to her firm's Westlaw subscription, and believed the service performed an internal citation check. None of that substitutes for opening the authority and reading it. She told the court she had relied on the tool without independently verifying every authority before filing, accepted responsibility and apologised, and now retrieves each decision through Westlaw, LexisNexis or another reliable source, checks quotations against the underlying opinions and runs a citation review before filing. Four things are verifiable from the record. The errors were caught by opposing counsel reviewing the insurer's motions, not by the AI tool or the firm's review process. They were material: motions in limine decide what a jury is allowed to hear. The victim of the fabrication was a homeowner, Fa'alagilagi Meni-Siliga, whose 2024 coverage suit over a Carson, California house damaged by fire and then storm water is still set for trial on October 5. And the failure mode is not confined to legal-specific tools - days earlier the New Mexico Supreme Court fined a defence attorney $5,000 over fabricated witness testimony and nonexistent material in a murder appeal that the lawyer attributed to ChatGPT. A legal-research product does not make the output safer; it moves the same next-token risk into a domain where a wrong result carries a court sanction. The check that works is the only one that ever worked: locate the case, read the quotation in the opinion, confirm it says what the filing claims it says. Source: https://www.lawcommentary.com/articles/state-farm-lawyer-ai-fake-cases-fine

Your answer

Sign in to verify this AI response.

Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.

More in Law

More from this topic

Unidentified 'legally trained' AI tool1 answer

I was representing myself in an unfair dismissal claim, so I used an AI tool billed as "legally trained" to prepare my submissions. It handed me case law to cite as precedent — specific legal principles and authorities — and it stated them as settled law, no hedging at all. At the hearing the Fair Work Commission said those authorities "may have been AI hallucinations and lacked any legal basis." How was I supposed to know the citations were not real?

AI voice-to-text court translation tool1 answer

I went to a California court counter for help with a restraining order, and the interpreter that day was an AI voice-to-text tool. It translated the clerk's instruction as: write down the name of the person "who receives the phone number, and address of the restraining order" — the clerk had actually said "who's getting the restraining order." In other pilot transcripts the same tool rendered "DUI" as "IUD" and "no contest" as "no competition." It didn't stumble or ask again, and it can't clarify who "he/she/they" means — it picks one subject and keeps going. Is that really what automated translation should mean when someone's legal rights are on the line?

Generative AI transcription tool1 answer

I ran a trial recording through an AI transcription tool and it handed back a clean, properly formatted transcript — every speaker labelled, nothing flagged as unclear or uncertain. It went in as the certified record. Then the Indiana Court of Appeals read it and found that a motion made by the prosecutor had been attributed to the trial judge, and an objection by the defendant had been attributed to the bailiff. The tool stated all of it as plain fact, with no hint that any speaker was a guess. Does a transcript like that still count as a true record on appeal?