Germany classifies Google AI Overviews, Perplexity as media providers

Germany's Commission for Licensing and Supervision (ZAK) announced its first rulings against AI services from Google and Perplexity, marking the first time German regulators have applied the country's State Media Treaty to AI search engines and chatbots. ZAK Chairman Dr. Thorsten Schmiege stated: "AI search engines and chatbots are content providers, and we are now consistently applying German media law to them." The rulings formally found that both companies violated Section 109 of the State Media Treaty and are immediately enforceable, with both firms given one month to appeal.

The legal classification hinges on a fundamental distinction: AI-generated responses are treated as the companies' own content, not neutral third-party distribution. Under the Digital Services Act, platforms distributing third-party content normally enjoy a liability shield that does not extend to a provider's own statements. Regulators found this shield inapplicable because the AI responses contain "independent, new, and substantive statements" produced by analyzing and combining material from various sources. A Munich court recently reached a similar conclusion and held Google liable for false claims in its AI-generated text; Google says it will appeal. The regulatory finding exposes Google and Perplexity to both civil liability and media law enforcement simultaneously.

Regulators specifically accused Google of failing to meet transparency rules and violating rules against discrimination. Google's AI summaries receive prime placement above traditional search results, pushing down links—particularly those to journalistic sources. Regulators view this as prohibited discrimination because Google's own AI content, not neutral search results, occupies the most visible position. Studies show that users rarely click source links once they feel their question has been answered, meaning even if links were moved higher in the layout, traffic to original sources would likely not recover substantially. Google says these studies are flawed but has not released data supporting an alternative view. Google does include source links within AI overviews, but the company's "Preferred Sources" feature—which allows users to choose which sources appear—has limited practical impact, as few users are likely to maintain a custom source list. Perplexity was flagged primarily for lacking a designated representative in Germany and missing transparency disclosures; regulators note the same concerns should apply in theory given the services' similar functioning, though Google's reach is far greater.

Regulators also frame AI services as media intermediaries. When a chatbot includes third-party content as sources or in link lists, it controls whether users can find that content through selection and placement. This meets the threshold for a media intermediary under German law, triggering transparency requirements meant to protect media diversity. Legal scholars Jan Oster and Christoph Busch have issued a related opinion supporting regulators' stance, arguing that adding AI to search engines fundamentally changes how people discover information—shifting from a linked list to a single prose answer—which cuts traffic to original sources and threatens journalism's funding. They recommend creating a separate regulatory category for AI search engines under state media law with explicit rules to protect media diversity.

Germany's ruling marks the first application of state media law to AI search engines and chatbots, establishing that AI-generated responses qualify as the providers' own content rather than neutral third-party distribution. This distinction is legally significant: it strips away the liability protections that platforms typically enjoy under the Digital Services Act, exposing Google and Perplexity to direct media law enforcement. The ZAK's reasoning rests on the observation that when AI services place their own summaries above traditional search results—especially journalistic sources—they act as intermediaries controlling content discovery. A Munich court recently reached a parallel conclusion, treating AI-generated text as independent content containing "independent, new, and substantive statements."

The enforcement reflects broader concern about traffic loss to original publishers. Studies cited in the ruling show that users rarely click source links once they feel their question has been answered by an AI summary. By elevating AI-generated prose above traditional search results, Google concentrates user attention on its own output rather than guiding them to journalistic sources. The legal opinion of Professors Jan Oster and Christoph Busch supports regulators' view that AI's role in search fundamentally changes how people discover information—shifting from a list of linked results to a single prose answer—which cuts traffic to original sources and threatens journalism's funding. Google has introduced a "Preferred Sources" feature allowing users to select which sources appear in AI responses, but regulators and legal scholars see this as insufficient protection for media diversity given the low likelihood that individual users will maintain custom source lists.