Skip to content

Generated Source Mirror

This page is generated from the source Markdown file so it can be viewed inside the MkDocs site.

📝 Page Status: Draft — sourced but not yet reviewed

  • Status: draft
  • Sources: 1
  • Relationships: 4
  • Research debt items: 3

Walters v. OpenAI, L.L.C.

Summary

Walters v. OpenAI, L.L.C., No. 23-A-04860-2 (Ga. Super. Ct., Gwinnett Cnty.), is a decided defamation action arising from a May 3, 2023 ChatGPT output that falsely described radio host Mark Walters as accused of embezzlement. On May 19, 2025, Judge Tracie Cason granted OpenAI summary judgment in full on three independent grounds: the output did not communicate defamatory meaning as a matter of law; Walters could not show fault; and he could not recover damages.

Verified Facts

  • Journalist Frederick Riehl used ChatGPT on May 3, 2023, while researching a Second Amendment Foundation suit against Washington’s attorney general. After ChatGPT accurately summarized pasted complaint text, it later produced an inaccurate summary accusing Walters of embezzlement when Riehl supplied a URL.1
  • Riehl had accepted terms warning that outputs can be incorrect, saw on-screen disclaimers, and had prior experience with fictional ChatGPT responses. He testified that within about an hour and a half he determined the output was untrue and that he did not believe it.1
  • Walters testified that he incurred no damages and did not ask OpenAI for a correction or retraction before suing.1
  • The court held that, in context — including contradictions, refusals, knowledge-cutoff warnings, and terms-of-use disclaimers — a reasonable reader in Riehl’s position could not have understood the output as stating actual facts, and that Riehl’s subjective disbelief independently defeated defamatory meaning.1
  • The court held Walters could not show negligence or, as a public figure, actual malice, and that he could not recover actual, presumed, or punitive damages on that record.1

Historical Context

Walters is among the first U.S. merits dispositions of a ChatGPT “hallucination” libel claim. It is a reputational-harm case, not a physical-injury product-liability suit, but it is the decided counterpart to the still-pending chatbot-harm inventory on Chatbot Product Liability and User-Harm Litigation.

Jurisdiction: Georgia Superior Court. Authority level: trial-court summary judgment. The holding is fact-bound: one reader, prominent disclaimers, rapid self-correction, and no damages. It does not decide that LLM output can never be defamatory, or that OpenAI can never be a publisher.1

The court’s public-figure and actual-malice analysis, and its use of product disclaimers to defeat defamatory meaning, are the pieces most likely to be cited in later hallucination suits. Those pieces remain Georgia trial-court law until an appellate court says otherwise.

Relationships

  • CASE-WALTERS-V-OPENAI cites SRC-WALTERS-V-OPENAI-SJ-2025.
  • CASE-WALTERS-V-OPENAI related_to TOPIC-AI-HARM-LITIGATION.
  • CASE-WALTERS-V-OPENAI related_to TOPIC-AI-LITIGATION.
  • CASE-WALTERS-V-OPENAI related_to TOPIC-FIRST-AMENDMENT.

Sources

  1. SRC-WALTERS-V-OPENAI-SJ-2025: Summary-judgment order (May 19, 2025).

Research Debt

  • Replace the lawrpg.org hosted PDF with a clerk-hosted or RECAP original.
  • Add any Georgia Court of Appeals or Supreme Court filing if Walters appealed.
  • Add the earlier federal removal/remand episode (N.D. Ga. / 11th Cir. fee dispute) only if OIR needs that procedural history.

Document metadata

  • Decision date: 2025-05-19
  • Last verified: 2026-09-03

  1. SRC-WALTERS-V-OPENAI-SJ-2025 — Summary-judgment order (May 19, 2025). Primary trial-court authority; not appellate precedent. ↩↩↩↩↩↩