A federal court just ruled that Google can face defamation lawsuits over the falsehoods its AI Overview spits out, piercing the Section 230 legal shield that Big Tech has hidden behind for decades while censoring Americans and dodging accountability for the content on their platforms.
The ruling matters because it establishes that AI-generated content — material produced by the platform itself, not merely passed through from users — isn't protected by Section 230. If platforms are legally responsible for what their AI invents, the entire framework shielding censorship-happy tech giants from consequences starts to crumble.
James Keene, a writer and TV producer whose prison experience alongside a suspected serial killer became the basis for the Apple TV+ miniseries "Black Bird," discovered last year that Google's AI Overview was telling anyone who searched his name that he had killed three women and was serving a life sentence. Keene's actual crime was marijuana-related. He served 17 months and was released. The AI had fabricated a triple homicide out of thin air.
According to court documents reviewed by the Mercury News, Keene "repeatedly alerted Google that the AI Overviews could be defamatory." The false statements stayed accessible for at least two months. Keene filed suit seeking at least $250 million in damages.
Google moved to dismiss. The company argued that AI Overviews don't count as statements of fact and that Keene, as a "limited public figure," would need to plausibly plead actual malice. The U.S. District Court in Illinois wasn't buying it — at least not entirely. The court denied two of Google's dismissal motions, granted two others, and allowed Keene to file an amended complaint. The case moves forward.
The stakes extend far beyond Keene. For years, platforms have leaned on Section 230 to moderate content with impunity — silencing dissenting voices, deplatforming users, and claiming neutral platform status when it suits them, all while their own algorithms and AI systems generate, amplify, and distort information. This ruling draws a line: when the machine itself invents the lie, the platform owns it.
The question now is how far that principle travels. If AI-generated hallucinations aren't shielded, what about algorithmic amplification, content curation, and the moderation decisions that have become the censorship infrastructure of the modern internet? The court opened a door. What walks through it next is the fight.
