German court holds Google liable for AI Mode's answers

The AI wrote it. The court says Google said it.

A regional court in Aachen has ordered Google to stop making false statements about a doctor in AI Mode, its AI-generated search feature. The order, dated Sept. 24 and reported on Sunday by The Decoder, treats the AI's answers as Google's own.

What the Aachen court ordered

The 8th civil chamber of the Aachen Regional Court issued a preliminary injunction against Google Ireland Limited in case 8 O 268/26. Because of the urgency, it ruled without hearing Google first.

The applicant is a physician who worked as a chief physician and professor of neuroradiology from 2010 until the end of June 2026. According to the order, AI Mode generated statements tying his departure to a criminal case. The court barred Google from making those statements about him in AI Mode. The Decoder reports the allegations concerned a different former professor.

The penalty for each breach:

Up to €250,000, or detention of up to six months.

Google also bears the costs of the proceedings.

Google was told, and declined

The timeline in the order is what makes this case sharp.

On Aug. 28, a colleague told the doctor what AI Mode was saying. The same day, he wrote to Google asking for the false statements to be deleted. On Sept. 15, Google told him it had decided to take no action. His lawyer sent a formal warning that day, and Google confirmed receipt on Sept. 16. The order says the AI-generated content was still available.

The court's reasoning is short. Google provides the AI-generated search, the chamber wrote, so it is responsible for the answers that search generates.

It added that checking was easily possible. Had Google put the doctor's objections to its own AI, the court said, it would have found the statements were false.

Two courts, one line

Aachen leaned on an earlier ruling. On May 28, 2026, the Munich Regional Court I held in case 26 O 869/26 that Google's AI does not just show search results. It makes new statements of its own, built from evaluating and combining third-party pages, which Google can check.

The Decoder reports that the Munich case involved two publishers that Google's AI falsely linked to fraud, and that Google announced an appeal but settled instead. So no higher court has ruled.

That matters. German search liability has long separated making a page findable from adopting its content as your own. Both courts put generated answers on the adopting side. A hallucination about a real person is now, in two German courtrooms, the operator's statement.

The Aachen order is not final. The Decoder notes Google can challenge it.

What it means if you ship AI that talks about people

If your product generates text about real people (a support bot, a research tool, a summary feature), read this as a German operating rule until a higher court says otherwise.

Four practical steps follow from the order's timeline:

• Give people a working channel to report false statements about them.
• Log what your system said, so you can see and reproduce the problem.
• Treat a credible complaint as notice. In Aachen, the ignored request did the damage.
• Have a way to stop a specific statement about a specific person, fast.

None of this needs a lawyer to start. It needs an owner.

What to watch

Watch whether Google files an objection in Aachen, and whether any case like this reaches a German higher regional court. Munich's settlement left that question open, and an appellate ruling would set the rule for everyone shipping generated answers in Germany.

Answer your takedown emails. In Aachen, that was the expensive part.