The Machine Invented a Crime and Pinned It on a Real Man. A Court Said That's Fine.
ChatGPT fabricated a fake lawsuit accusing radio host Mark Walters of embezzling from a gun-rights nonprofit he never worked for. He brought the first defamation suit against a generative AI. The court threw it out. Because everyone already knows the machine lies.
What the machine said
In 2023 a journalist named Fred Riehl asked ChatGPT to summarize a real lawsuit. ChatGPT obliged by summarizing a different one, one that did not exist. In it, Mark Walters, host of the syndicated Armed America Radio, had embezzled funds from the Second Amendment Foundation.
Walters had never worked for the Second Amendment Foundation. There was no embezzlement. There was no case. The model produced a complete, confident, fictional accusation of a financial crime, attached it to a named, real, living person, and served it on request.
The first AI-defamation suit
Walters sued OpenAI. The first defamation suit against a generative AI product. For a moment it looked like the question would finally get answered: when a machine fabricates a crime about you, who is liable?
The answer, and why it is worse than no answer
In May 2025, a Georgia court granted OpenAI summary judgment. The output was not defamatory as a matter of law, the court held, because no reasonable reader would have taken it for a statement of fact. The journalist knew ChatGPT fabricates. He had seen the warnings. He could be expected to check.
Read what that means. The legal system’s answer to “an AI invented a crime and attached my name to it” is: that’s fine, because everyone already knows the machine lies.
The defense against an AI that defames you is that nobody should have believed it in the first place. That is not a safeguard. It is a confession entered as an exoneration — the product is so unreliable that its lies are legally weightless, and that unreliability is the thing that protects the company that sells it.
The disinformation engine does not need a human operator pointing it at a target. It will fabricate the target, the crime, and the citation on its own. And the law, for now, has decided that the machine’s known unreliability is everyone’s problem except the maker’s.
Sources (gate-verified): Eric Goldman, Technology & Marketing Law Blog — ChatGPT Defeats Defamation Lawsuit Over Hallucination (Walters v. OpenAI) · Cleary Gottlieb — Georgia Court Dismisses Defamation Lawsuit Against OpenAI Over ChatGPT Output