German Court Finds Google Liable After AI Overview Falsely Labels Companies as Scams

A German court ruled against Google after its AI Overview feature confidently told users that two publishers were scams with a history of fraud — a claim the AI fabricated entirely. Google's defense rested on the small-print disclaimer at the bottom of its search results: "AI can make mistakes, so double-check responses." The court was not persuaded that a boilerplate caveat absolves a company of responsibility for defamatory hallucinations, and Google is already planning an appeal.

The case highlights a pattern the article's author calls a "dilemma": tech companies selectively argue that their AI either is or is not like a responsible agent, depending on whichever framing helps them dodge liability in a given lawsuit. Google argued its AI merely surfaces others' content (not responsible); Character Technologies argued its chatbot's outputs were free speech (responsible, and thus rights-bearing). Courts have so far rejected both maneuvers, leaving companies in a bind: acknowledge the AI as a creative actor and accept accountability, or admit it's a dumb aggregator and lose the "transformative fair use" defense on copyright too.