When we do a search on Google, the first thing that appears is an AI-generated summary related to that search. This summary, known as the AI Overview, also typically shares a series of links so we can see where you got the information. But what if there are no links? Who is to blame when Google’s AI claims something without those links as support?
A court in Aachen, Germany is clear: Google. And it ordered the company to prevent its AI mode from spreading false claims about a doctor whom Google’s AI summary accused of sexually abusing his secretary. Crime exists, but it has been attributed to the wrong person. Now the court, relying directly on a ruling handed down in Munich in May, considers Google’s AI responses as the company’s own statements. This is a precautionary measure at the moment and not yet final, so Google can still appeal
the case According to legal media Beck-Actuelle, the plaintiff is a doctor who worked as chief doctor at a German hospital until last summer. In late August, a colleague told him that when he searched for his name in Google’s artificial intelligence mode with the word “discount,” a story about him came up that was quite serious and alarmingly broad in tone. It alleged he harassed his secretary on a business trip, was fired, received a suspended sentence for sexual assault and walked away with a six-figure severance package.
None of this was true. And according to Beck-Actuelle, the incidents appear to relate to another former member of the management staff at the same clinic, and the news articles linked by Google as sources never mentioned the plaintiff. According to The Decoder, additional searches with terms like “termination” and “judicial proceedings” also resulted in equally detailed misinformation.
Google’s response. The doctor asked Google via email to remove the content by the end of August. However, Beck-Actuelle said Google responded that it had reviewed the complaint and would not take any action. A warning letter from his lawyers, the Cologne law firm Hawker, sent the same day, also changed nothing.
What did the court decide? The Aachen court ruled that AI-generated texts violated the doctor’s general right to personality, as they attributed criminal charges and employment outcomes that affected another person. According to the media, he issued the order in an emergency manner without first consulting Google, alleging damage to his professional and social reputation. Google faces fines of up to 250,000 euros for each violation.
The Munich example. In an earlier case, the Munich Regional Court I ruled that Google could be directly liable for its AI content, as shared by DW. The plaintiffs were two Munich publishers who falsely linked the feature to subscription scams and fraudulent schemes. Google argued that users can check sources through links and the AI doesn’t know how to blindly trust the results. The court rejected this argument, holding that the summary was a self-contained statement and gave no indication to the reader that it might be unreliable. Google announced at the time that it would appeal the ruling.
Decoder Media reports that the appeal was never resolved, since, according to the investigation, the case ended in an out-of-court settlement, and neither Google nor the publisher wanted to clarify whether there was financial compensation or why the case was closed.
Cover image Shutter speed
In Xataka | Anthropology shows AI’s impact on junior employment: 27% to 9% of new hires
