The ChatGPT Lawsuit: Walters v. OpenAI and What Brands Need to Know

Originally published July 2023. Updated June 2026.

ChatGPT was the first major language model to face a U.S. defamation lawsuit. In June 2023, Georgia radio host Mark Walters sued OpenAI in Gwinnett County Superior Court after the chatbot generated a false legal summary describing him as having embezzled funds from a non-profit. The accusation was never made against Walters in any real-world case. ChatGPT had fabricated it.

The case was the first of a category. Three years on, the questions it raised — about liability, hallucination, and what Section 230 protects — define how communications, legal, and brand teams now operate around the answer engines. This piece sits inside the broader ChatGPT communications profile, which covers the platform itself, the citation share it controls, and the discipline of building a brand against it.

What Walters v. OpenAI Was About

Journalist Fred Riehl asked ChatGPT to summarize a real federal lawsuit. The chatbot returned a fabricated summary that named Walters, accused him of fraud, and inserted detail that did not exist in the underlying case. Walters discovered the false summary, filed for damages, and put the question of AI defamation liability into a U.S. courtroom for the first time.

OpenAI's position was that ChatGPT's homepage disclaimer — the system may produce incorrect information — limited the company's exposure. The court ultimately dismissed the case in 2024 on the grounds that Walters had not shown actual malice or actual damages. The dismissal did not, however, resolve the underlying liability question. It deferred it.

Why the Hallucination Problem Has Not Gone Away

The technical issue at the center of Walters is unresolved. Large language models do not retrieve facts. They predict the next token. When the prediction lines up with reality, the answer is correct. When it does not, the model produces a hallucination — fluent, confident, and wrong. This is the discipline of answer engines brand and communications teams have to learn.

Hallucinations now show up across every major model — ChatGPT, Claude, Gemini, Perplexity. The frequency varies. The risk profile does not. A model that fabricates an embezzlement allegation against a radio host can fabricate one against a CEO, a brand, a charity, or a candidate.

Section 230 and AI

Section 230 of the Communications Decency Act gives U.S. internet companies immunity from liability for third-party content posted on their platforms. The question Walters raised — and the courts have not finally answered — is whether content generated by an AI system counts as third-party content at all.

If the model produces the content, the model's operator is the publisher. That theory, if upheld, would unwind Section 230 protection for AI outputs and open every model operator to defamation, copyright, and privacy liability for what their systems generate. Courts are working through it. Brands and communications teams should not wait for the answer.

What This Means for Brands and Communications

The Walters case crystallized three operational realities every crisis communications team now plans around.

One — your brand can be defamed by a chatbot. Customer service complaints, executive misconduct allegations, product safety claims — all can be hallucinated into existence by an AI system answering an unrelated question. The defamation does not need a publisher in the traditional sense. The chatbox produced it.

Two — the 48-hour breach window applies. When a hallucinated claim about a brand surfaces in ChatGPT or Perplexity, the corrective response has to happen fast — through earned media, owned content, and structured corrections to the underlying citation footprint.

Three — citation share is the defense. The brands that own the high-quality, easily retrievable information about themselves in the engines are the brands least vulnerable to AI-generated fabrication. The discipline is AI Communications — building authority inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews before the next hallucinated claim surfaces.

The Legal Landscape in 2026

Since Walters, additional AI defamation and copyright cases have moved through U.S. and European courts. The New York Times v. OpenAI litigation continues. Multiple authors and publishers have filed against OpenAI, Anthropic, and others on training-data questions. The EU AI Act came into force. None of these cases have produced the definitive ruling on AI defamation that Walters could have set.

The unresolved question matters. Until U.S. courts decide whether AI operators are publishers or platforms, every brand operates in a gray zone — exposed to hallucination, unclear on recourse, and reliant on its own communications infrastructure to shape the answer.

Related Reading on Everything-PR

• ChatGPT: The Communications Profile of the AI Era — the full entity hub: platform, usage, competitive landscape, and what every brand needs to know.
• Hallucination — the technical definition of the failure mode at the center of this case.

Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Thirty-plus publications. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.