Germany

[DE] Media regulator applies media law rules to certain AI content for the first time

IRIS 2026-8:1/22

Christina Meese

Institute of European Media Law

On 14 July 2026, the Commission for Licensing and Supervision (Kommission für Zulassung und Aufsicht – ZAK), the central body of the 14 German state media authorities, announced that, for the first time, German media law could apply to AI searches and AI chatbots. Its decisions regarding Google's AI Overviews and Perplexity's AI chatbot of the same name, which features an AI news page, indicate that the provisions of the Inter-State Media Treaty (Medienstaatsvertrag – MStV), including those safeguarding diversity, will be applied consistently in future. Providers will not be able to invoke the liability exemptions under the EU Digital Services Act (DSA), or will only be able to do so to a limited extent, because AI-generated content will, in part, be classified as their own content rather than third-party content.

The ZAK is the central body for cooperation between the heads of the 14 state media authorities. It is responsible for the licensing and supervision of nationwide private broadcasters, online media and platforms on the basis of the relevant provisions of the MStV. Its decisions were issued in two proceedings conducted by the media authorities of Hamburg-Schleswig Holstein and Berlin-Brandenburg. Amongst other things, they established, partly as a necessary preliminary matter and basis for any further proceedings, that the provisions of the MStV also apply to AI-generated content. In this regard, the ZAK relied primarily on a recently published legal opinion it had commissioned on the integration of AI applications into search engines, which concludes, in essence, that AI-generated responses are, in principle, generally to be classified as the provider's own content under media and freedom of expression law. According to this legal opinion, this applies in particular to "hallucinated" content; however, even when existing sources are processed, the repackaging, condensation and blending of content suggest that AI output is independent. A different conclusion may be reached if it is clearly apparent to users that only third-party content is being reproduced. Consequently, both the legal opinion and the ZAK conclude that operators are, in principle, liable for AI responses under general law, and that the limitations on liability set out in Articles 4 to 6 of the DSA do not apply because they are linked to "information provided by a user".

The ZAK's decisions also take into account this distinction between AI’s own responses and third-party content, particularly journalistic sources to which AI Overviews and Perplexity regularly refer. The services are to be classified as media intermediaries within the meaning of the MStV, i.e. as telemedia that also aggregate, select and present content from third-party journalistic and editorial sources in a manner that is accessible to the general public, without combining it into a single comprehensive service. Under Sections 93 and 94 of the MStV, such media intermediaries are subject to transparency and anti-discrimination rules in relation to the journalistic content they present. The ZAK considers that these rules are not (fully) being complied with, or at any rate intends to monitor compliance with them in future. In AI Overviews, for example, AI responses are presented as a significant part of the results and appear clearly visible and prominently placed. Consequently, the traditional list of links in Google Search is harder to find than the AI responses – which are considered separate content – and is thus unfairly discriminated against. Perplexity, as an AI chatbot, on the other hand, attaches third-party content as "sources", "further information" or in the form of entire lists of links to AI-generated responses, thereby also determining its discoverability. In this respect, too, the provider must comply with the ban on unjustified unequal treatment of journalistic sources.