What happened

  • Counsel for plaintiff filed two notices of errata to correct errors in a memorandum of law. That is how the court found the problem: it looked at the corrections and discovered the memorandum "contained at least two fake citations."
  • One citation was to a case that did not exist. The other cited a case that "did not include the quotation attributed to it." The order does not name either authority.
  • The case settled shortly afterward. The court scheduled a hearing anyway to investigate further, and it was held on October 1, 2026.
  • At the hearing, counsel explained that in his haste to meet a filing deadline he did not properly review the document, which included materials from research done with AI. The order does not name the AI tool.
  • Before the hearing, counsel had already told and apologized to his client and to opposing counsel (who said they were not prejudiced), given his entire fee (about $10,000) to Jacksonville Area Legal Aid, completed 7.5 hours of CLE on AI ethics, and described changes to his office procedures.

What the court said

The court treated counsel as candid and did not escalate. It found that he "understands the gravity of his actions and will not repeat them," and that the measures he had already taken were "a suitable remedy." No further action was ordered, and the file remains closed.

The court also repeated language from an earlier order in another of its cases, Hodges v. Meridian Waste Acquisitions, LLC: lawyers are ethically bound to give "true and accurate" citations, and AI use "does not alter this basic requirement of candor to the tribunal." It added that relying on AI without verification "is not acceptable conduct; it will not be tolerated."

What it means for your firm

  • The court's discovery path was mundane. An errata filing led to a cite-check by the court, and a hearing followed even though the case had settled. Settlement did not end the court's interest.
  • The outcome was lenient because of what counsel did before the hearing: prompt disclosure, repayment of the fee, training and process changes. Our reading: that sequence is a usable template for how to respond if a bad citation surfaces in your own filing.
  • Filing deadlines were the stated cause. Our reading: the exposure is highest when AI-assisted research reaches a brief in the last hours before a deadline, with no second reader.
  • The court did not treat AI as a mitigating factor. The duty of accurate citations stayed with the lawyer who signed.

What to do Monday

  1. Require a cite-check of every authority and every quotation, against the source, before any memorandum is filed, regardless of how the research was done.
  2. Set an internal cutoff before each filing deadline so the cite-check happens on a finished draft and not in the final hour.
  3. Write down what you will do if you find a bad citation after filing: who corrects the record, who tells the client and opposing counsel, and how fast.
  4. Put AI-use disclosure to the supervising lawyer into your filing checklist, so the signer knows which sections came from AI-assisted research.

Sources

Questions and answers

Did the court sanction the lawyer in Glenn v. Absolute Resolutions?

No. The court found that counsel's apology, notifications, fee donation, CLE and office changes were "a suitable remedy" and said no further action would be necessary.

How did the court find the fake citations?

Counsel filed two notices of errata to correct errors in the memorandum. In reviewing them, the court discovered at least two fake citations.

Does the order say which AI tool was used?

No. The order says only that the memorandum "included materials generated by research conducted using AI." It does not name the tool, the nonexistent case, or the case that was misquoted.

Operational information, not legal advice. Facts come from the court record linked under Sources, read on Oct 21, 2026. The reading of what it means for your firm is ours. A source link does not replace the full docket or later history.