Keene v. Google: A Libel Case Tests AI-Generated Search Results
The ruling rejects Google's argument that its AI summaries are mere starting points, finding they can be read as factual assertions.2
Keene, whose story inspired an Apple TV series, seeks $250 million, claiming the false statements persisted for months after he notified the company.1
- A federal judge ruled a defamation lawsuit against Google, centered on its AI Overview feature, can proceed, marking an early legal test for liability over AI-generated search summaries.1
- Plaintiff James Keene alleges the AI falsely stated he was convicted of murdering three women and serving a life sentence, while his actual conviction was marijuana-related.1,2
- The court found Keene plausibly alleged Google acted with 'actual malice,' citing his repeated notifications to the company and its alleged continued publication of the statements.2
- Google argued the AI Overviews are not statements of fact and that Keene, as a limited public figure, must prove malice, arguments the court considered but found insufficient to dismiss the case at this stage.1
- Material facts remain unresolved, including the exact text of the AI statements, the content of Keene's notifications to Google, and the basis for his $250 million damages claim.1
A U.S. District Court in Illinois ruled that James Keene may proceed with a lawsuit alleging Google's AI Overview feature defamed him by falsely stating he was a convicted murderer serving a life sentence.1
Judge Thomas Durkin allowed claims related to three specific AI Overview responses to advance, dismissing a fourth claim but permitting Keene to amend his complaint.2
The court held that an ordinary reader would understand the AI's statements as factual assertions, rejecting Google's argument that they were not statements of fact.2
The Allegations: AI Overviews and a Criminal Record
James Keene alleges that Google's AI Overview feature generated false and defamatory statements about his criminal history.1
Keene says the AI stated he had killed three women and was serving a life prison term.1
His actual conviction, according to his allegations, was marijuana-related, for which he served 17 months.1
In his complaint, Keene alleges the AI Overviews falsely stated he was serving a life sentence without parole for multiple convictions, for the murders of three women, and was convicted of drug trafficking.2
Keene further alleges the AI statements cited sources that did not contain the false information.2
The Court's Ruling on Google's Motion to Dismiss
Judge Thomas Durkin of the U.S. District Court for the Northern District of Illinois ruled on Google's motion to dismiss the case.2
The court allowed the lawsuit to proceed on claims related to three specific AI Overview responses, identified in the litigation as Overviews 1, 2, and 4.2
Judge Durkin dismissed the claim regarding a fourth AI Overview, identified as Overview 4, for failure to plausibly allege publication to a third party, but granted Keene leave to file an amended complaint on that point.2
Key Legal Findings at the Pleading Stage
The court rejected Google's argument that its AI Overviews are not statements of fact, holding instead that an ordinary reader would understand them as factual assertions.2
Judge Durkin also rejected Google's contention that the summaries were merely a 'starting point' for research, finding this defense insufficient to dismiss the case at the pleading stage.2
On the critical issue of 'actual malice'—a requirement for defamation claims brought by public figures—the court found Keene's allegations sufficient to proceed.2
The ruling states Keene satisfied this requirement by alleging he notified Google three times of the falsity and that Google acknowledged or apologized yet continued publishing the statements.2
Judge Durkin found the emails described in the complaint plausibly supported an inference of actual knowledge or reckless disregard for the truth.2
The Parties' Positions and Damages Claim
Google argued in its motion to dismiss that AI Overviews are not statements of fact and that Keene, whom it characterized as a 'limited public figure,' must plausibly plead malice.1
Keene's lawsuit seeks at least $250 million in damages.1
Keene alleges he 'repeatedly alerted Google that the AI Overviews could be defamatory' and that the false statements were accessible for at least two months.1
What Remains Unresolved
The exact text of the alleged AI Overview statements (1, 2, and 4) that the court is allowing to proceed has not been made public in the available evidence.2
The specific content and dates of the emails or notifications Keene sent to Google, which the court cited for the actual malice pleading, are not detailed in the reports.2
The full text of Judge Durkin's court order is not provided, leaving the precise legal reasoning and scope to be verified.2
Google's full motion to dismiss arguments and the court's complete ruling on each point, beyond the summaries provided, are not available.1
The basis for the $250 million damages figure—how it was calculated or pleaded—is not established.1
It is not shown whether the sources cited by the AI Overviews contained the alleged false information at the relevant time.2
Keene alleges the false AI Overview statements about him were accessible for at least two months.1
Keene alleges he notified Google three times about the defamatory AI Overviews.2
Reason reports that Judge Thomas Durkin allowed the libel-by-AI claim to proceed.2
Mercury News reports that a U.S. District Court in Illinois ruled that James Keene may sue Google for defamation over AI Overview responses.1
- The exact text of the alleged AI Overview statements (1, 2, and 4) that the court is allowing to proceed.
- The specific content and dates of the emails or notifications Keene sent to Google, which the court cited for the actual malice pleading.
- The full text of the court's order (Judge Durkin's memorandum opinion and order) to verify the precise legal reasoning and scope.
- Google's full motion to dismiss arguments and the court's complete ruling on each (beyond the summary provided).
- The basis for the $250 million damages figure (how it was calculated or pleaded).
- Whether Keene's Wikipedia page(s) cited by the AI Overviews have been archived to show the alleged discrepancy at the relevant time.
- Whether the sources cited by the AI Overviews contained the alleged false information at the relevant time.
- 1.
- 2.