Walters v OpenAI LLC
AI-generated text. This page was generated using artificial intelligence.
Walters v. OpenAI, L.L.C. is a Georgia defamation case concerning false statements generated by ChatGPT about radio host Mark Walters. Walters filed suit on June 5, 2023. On May 19, 2025, Judge Tracie Cason granted summary judgment in full and entered judgment for OpenAI.[3][1]
Claim and factual record
Walters alleged that ChatGPT, responding to journalist Frederick Riehl’s request to summarise an unrelated lawsuit involving the Second Amendment Foundation, falsely described Walters as accused of embezzlement. Walters was not a party to that lawsuit. He sought general and punitive damages, litigation costs and attorney fees.[3]
The summary-judgment order describes Riehl’s May 3, 2023 interaction with ChatGPT. It recounts accuracy warnings, a warning that ChatGPT could not retrieve the linked complaint, contradictory responses, and Riehl’s prior experience with fictional outputs. The court found that Riehl did not believe the challenged statements and confirmed their falsity shortly afterwards.[1]
May 19, 2025 summary judgment
The court identified three independent grounds for judgment:[1]
- Defamatory meaning. In the specific context, a reasonable reader in Riehl’s position would not understand the challenged output as communicating actual facts. Riehl also did not himself believe it.[1]
- Fault. Walters had not produced evidence establishing ordinary negligence. The court also treated him as a public figure, or at least a limited-purpose public figure for the relevant controversy, and found no evidence meeting the actual-malice standard. General knowledge that ChatGPT can make mistakes did not establish knowledge or reckless disregard of the particular statements’ falsity.[1]
- Damages. Walters conceded that he had no actual damages. The court found punitive damages barred by his failure to request a correction or retraction and rejected presumed damages on the record. It also held that presumed or punitive damages for these statements on a matter of public concern required actual malice, which had not been shown.[1]
The ruling applies defamation standards to the particular output, recipient and evidentiary record; it does not establish that every false AI-generated statement is immune from liability.[1]
Earlier federal removal and fee appeal
OpenAI removed the case to federal court and later withdrew its removal notice. In an April 1, 2024 unpublished decision, the Eleventh Circuit vacated the unexplained denial of Walters’ request for removal-related costs and attorney fees and remanded for an explanation. That appeal concerned fees following removal, rather than the later state-court defamation judgment.[2]
Related articles
References
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 Superior Court of Gwinnett County, Walters v. OpenAI, L.L.C., No. 23-A-04860-2, order granting summary judgment, filed May 19, 2025, pp. 1–22; filed court order reproduced by Reuters.
- ↑ 2.0 2.1 Eleventh Circuit, Mark Walters v. OpenAI, L.L.C., No. 23-13843, unpublished opinion, April 1, 2024, pp. 1–3.
- ↑ 3.0 3.1 3.2 3.3 3.4 3.5 Walters v. OpenAI, L.L.C., complaint, court-stamped June 5, 2023, No. 23-A-04860-2, pp. 1–5 and attached exhibit; court filing reproduced by Ars Technica.