Google tin beryllium straight liable for mendacious claims successful AI Overviews, a German tribunal ruled. The Regional Court of Munich recovered that AI-generated summaries are Google’s ain contented and not protected arsenic accepted hunt results are.
The tribunal issued a impermanent injunction barring Google from repeating mendacious claims astir 2 Munich publishers, according to The Decoder. This lawsuit centered astir AI Overviews wrongly linking the publishers to scams, subscription traps, and questionable business practices.
AI Overviews aren‘t hunt results. AI Overviews do much than thief users find third-party content, the tribunal said. They rewrite, combine, and measure accusation “in its ain words and according to its ain structure.”
- In the disputed searches, Google’s AI Overview allegedly presented standalone claims astir questionable business practices, on pinch warnings and reddish flags. The tribunal recovered those claims didn’t look successful the linked sources.
- Because Google created the feature, controls its presentation, and controls the underlying algorithms, the tribunal treated the statements arsenic Google’s ain content.
Search protections didn’t apply. Google based on that German lawsuit rule limiting liability for accepted hunt engines and autocomplete should apply. Those rules mostly dainty hunt providers arsenic indirect infringers erstwhile they aboveground third-party content.
- The tribunal rejected that argument. Unlike accepted hunt results, which nonstop users to outer pages, AI Overviews make caller substantive statements from aggregate sources.
- The tribunal besides rejected Google’s declare that users could verify the accusation by reviewing the cited links. It recovered the AI Overview stood connected its ain and presented its claims arsenic a complete answer.
Why we care. The ruling doesn’t dainty AI Overviews for illustration neutral hunt links. If an AI-generated summary makes mendacious claims astir a company, Google whitethorn beryllium straight liable for those statements.
Wrong links, incorrect claims. The tribunal recovered that the AI Overview conflated accusation astir different companies pinch that of the 2 publishers and created unsupported connections to the root material.
- Because the linked sites didn’t make the disputed claims, the publishers would person had nary clear 3rd statement to writer if Google were treated only arsenic a hunt intermediary.
- The tribunal said Google tin comparison AI-generated statements pinch underlying sources, astatine slightest successful cases for illustration this.
Google must extremity repeating the claims. The injunction bars Google from repeating astir of the challenged claims, including allegations involving scams, subscription traps, questionable institution ties, telephone calls that allegedly ne'er occurred, and deficiency of availability.
- Google must salary 80% of the ineligible costs. Each patient must salary 10%.
- The tribunal said the consequence of repetition violations remains because Google did not supply a cease-and-desist declaration pinch a punishment clause, and the aforesaid algorithms could make akin claims again.
- The ruling is impermanent and whitethorn still beryllium challenged. However, it gives publishers and brands a way to situation mendacious AI Overviews arsenic Google’s ain statements alternatively than simply hunt results.
English (US) · 
Indonesian (ID) ·