State v. Coleman
Ohio Court of Appeals, Eleventh Appellate District (Ashtabula County) · Ohio Ct. App. (11th Dist.) · Ohio bar guidance
Verified May 5, 2026
- Citation
- State v. Coleman, 2026-Ohio-965, No. 2024-A-0040 (Ohio Ct. App. 11th Dist. Mar. 20, 2026)
- Decided
- March 20, 2026
Summary
Attorney William B. Norman filed an application to reopen the criminal appeal of Malikhi Jermaine Coleman that attributed an inflammatory quotation to the trial prosecutor, citing trial transcript page 559. The cited page was in fact the court reporter's signature page, and the quotation did not exist anywhere in the record. Norman's paralegal had drafted the application using ChatGPT, which fabricated the quotation, and Norman filed it under a sworn affidavit without verifying the quotations against the transcript. Judge Eugene A. Lucci, joined by Judge Robert J. Patton, imposed sanctions after a show-cause hearing; Judge John J. Eklund concurred in part and dissented in part.
- AI tool:
- ChatGPT
- Sanction amount:
- $2,000
What sanction did the court impose?
$2,000 monetary sanction (credited against an earlier settlement payment to the prosecutor's office), referral to the Ohio Office of Disciplinary Counsel, removal as counsel for Mr. Coleman, striking of the application to reopen, six hours of mandatory CLE on AI ethics, two-year certification requirement on every Ohio filing, two-year obligation to serve the order on the presiding judge of every new court of appearance, and written letters of apology to the trial prosecutor, trial judge, trial counsel, and prior appellate counsel.
Why does State v. Coleman matter for law firms using AI?
State v. Coleman is among the most expansive sanctions orders yet issued for AI-generated fabrications in a court filing, and the first reported Ohio appellate decision of its kind. The Eleventh District found that Norman’s misconduct went beyond a single hallucinated quotation: he failed to correct the record after the prosecution flagged the fabrication, appealed to the Supreme Court of Ohio without disclosing the infirmity, and two months after the sanctions hearing filed a motion in another case that contained an embedded ChatGPT prompt. For managing partners, the case is a reminder that supervisory liability under Prof.Cond.R. 5.3 attaches when a paralegal uses a public AI tool, that a remedial AI policy proffered to a court will itself be scrutinized, and that the duty to correct a tribunal is personal to the signing attorney and is not discharged by an adversary’s identification of the falsehood.
Sources
Primary sources
Further reading
Source PDF is a Westlaw printout mirrored from the Damien Charlotin hallucination database. We are working to add the underlying court docket (PACER, CourtListener, or court website) as a second source.