How the Digital Interstate Media Treaty regulates media, platforms and AI

With the Digital Interstate Media Treaty (Digitale Medien-Staatsvertrag), the German states intend to adapt the Interstate Media Treaty (Medienstaatsvertrag, MStV) to new European requirements for digital media, political advertising and artificial intelligence. The draft of the first part (Ninth Interstate Media Amendment Treaty, version of 10 June 2026) mainly affects media companies, online platforms, providers of journalistic content and the state media authorities (Landesmedienanstalten) as supervisory authorities. The heads of government of the states approved the draft on 25 June 2026 and intend to sign it by January 2027 at the latest. It is scheduled to enter into force on 1 August 2027; if not all instruments of ratification have been deposited by 31 July 2027, it will become void.

The Interstate Media Treaty primarily contains the provisions required to give effect to the European Media Freedom Act (EMFA), the Regulation on the transparency and targeting of political advertising and the AI Act. The Regulations are, in principle, directly applicable; the Interstate Media Treaty mainly determines who monitors compliance with them. For companies, this means: new transparency, cooperation and compliance obligations meet extended supervisory powers of the state media authorities. For the first time, the draft also brings the press within the scope of the Interstate Media Treaty and defines what is to be regarded as a print product (Druckerzeugnis) and as a periodical print product (periodisches Druckerzeugnis) (Section 1(1) and (2), Section 2(2) nos. 32 and 33 MStV-E; MStV-E refers to the Interstate Media Treaty as amended by the draft).

New competences regarding AI

The state media authorities are given an express role as market surveillance authorities for AI applications in the media sector.

This competence builds on the federal AI Market Surveillance and Innovation Promotion Act (KI-Marktüberwachungs- und Innovationsförderungsgesetz, KI-MIG), which entered into force on 29 July 2026. Under Section 2(1) KI-MIG, the Federal Network Agency (Bundesnetzagentur) is the market surveillance authority for the AI Act unless the Act provides otherwise. Section 2(8) KI-MIG provides for an exception for media service providers within the meaning of Art. 2(2) EMFA that provide, take into service or use AI systems in operating, offering, distributing and making available media services for journalistic or advertising purposes. For these cases, the federal act leaves market surveillance to the authorities competent under the law of the states. In this area, the Federal Network Agency merely receives complaints and forwards them to the competent authority (Section 8 KI-MIG). The new Section 111(5) MStV-E builds on Section 2(8) KI-MIG and designates the state media authority competent under Section 106 MStV as the market surveillance authority.

The state media authorities are therefore competent only where two conditions are met. The first concerns the provider: only media service providers within the meaning of the EMFA are covered, i.e. providers whose professional activity is to provide a media service, who have editorial responsibility for the choice of the content and who determine the manner in which it is organised. These include private television and radio broadcasters, providers of on-demand services, press publishers, journalistic online services and the public service broadcasters. Online platforms are covered only where, exceptionally, they themselves have editorial responsibility for the choice of their content. Search engines do not fall within the EMFA definition, and providers of AI services such as chatbots are, as a rule, not media service providers because they lack editorial responsibility. For them, the Federal Network Agency remains the market surveillance authority; for general-purpose AI models, it is the European Commission’s AI Office. The second condition concerns the purpose: the AI system must be used for journalistic or advertising purposes. If a publisher uses AI e.g. in recruitment or accounting, the Federal Network Agency remains competent. Depending on the specific use cases, the same media company may therefore be subject to two supervisory authorities.

Section 111(5) MStV-E does not limit supervision to individual provisions of the AI Act. The state media authorities monitor all obligations incumbent on the media service provider as provider or deployer of an AI system; an earlier draft had limited their competence to the enforcement of Art. 50 AI Act. In practice, supervision will mainly concern the transparency obligations under Art. 50 AI Act. Anyone deploying AI to generate or manipulate image, audio or video content constituting a deep fake must disclose this; where the content forms part of an evidently artistic, satirical or fictional work, disclosure in an appropriate manner that does not hamper the display or enjoyment of the work is sufficient (Art. 50(4), first subparagraph, AI Act). The scope of the term “deep fake” is disputed. On 5 June 2026, the Broadcasting Commission of the states (Rundfunkkommission) criticised the broad interpretation in the European Commission’s draft guidelines on Art. 50 AI Act, arguing that press, audio and video content published under editorial responsibility would otherwise be subject to disproportionate labelling obligations. AI-generated text published with the purpose of informing the public on matters of public interest must also be disclosed. This obligation does not apply where the text has undergone a process of human review or editorial control and a person or company holds editorial responsibility for their publication (Art. 50(4), second subparagraph, AI Act). As a rule, the disclosure obligation therefore does not apply to texts published under editorial responsibility; the AI Act provides for no comparable exception for image, audio and video content. Where a media house operates its own AI applications, such as a chatbot on its website, the provider obligations under Art. 50(1) and (2) AI Act apply in addition: human beings must be informed that they are interacting with an AI system, and synthetic content must be marked in a machine-readable format.

Fines for infringements of the AI Act are imposed by the respective market surveillance authority (Section 17(1) KI-MIG), which will in future be the state media authority in the media sector. For infringements of Art. 50 AI Act, Art. 99(4) AI Act provides for administrative fines of up to EUR 15 million or up to 3 % of total worldwide annual turnover for the preceding financial year, whichever is higher. The state media authority may initiate proceedings against public broadcasters and against providers of periodical print products only after consulting the competent supervisory body or the German Press Council (Deutscher Presserat), respectively. Their opinion is binding, provided that the limits of the margin of appreciation have been observed (Section 111(6) MStV-E). For cooperation with the Federal Network Agency, the Digital Services Coordinator and the European Commission, the state media authorities appoint a joint representative.

The period until the Interstate Media Treaty enters into force remains unresolved. The transparency obligations under Art. 50 AI Act have applied since 2 August 2026. Since then, Section 2(8) KI-MIG has assigned market surveillance in the media sector to the authorities competent under the law of the states, but does not itself designate any authority; the Interstate Media Treaty is to designate the state media authorities only with effect from 1 August 2027. Supervision under the Interstate Media Treaty continues irrespective of this: on 14 July 2026, the Commission on Licensing and Supervision (ZAK) for the first time issued decisions against AI services of Google and Perplexity. However, the Interstate Media Treaty does not confer the powers of a market surveillance authority under the AI Act, since Section 109 MStV requires an infringement of the Interstate Media Treaty itself. Neither the KI-MIG nor the Interstate Media Treaty specifies which authority may pursue infringements of Art. 50 AI Act by media service providers until August 2027.

Competences and new sanctions under the EMFA

To enforce the European Media Freedom Act, the draft extends the catalogue of administrative offences. Infringements of certain obligations, for example relating to functionalities of online platforms, transparency information, dialogue procedures with media service providers or the visibility of media offerings on user interfaces, may be prosecuted as administrative offences (Section 115(1a) MStV-E). The fine may amount to up to EUR 1.5 million (Section 115(2) MStV-E); the limitation period for prosecution will in future expire only after 24 months instead of six months (Section 115(5) MStV-E).

These obligations include, for example, Art. 18 EMFA, which requires providers of very large online platforms such as Facebook, Instagram, TikTok and YouTube to provide media service providers that have declared themselves as such to the platform with a statement of reasons before suspending the provision of their services in relation to that content or restricting its visibility because the content is incompatible with the platform’s terms and conditions, to give them the opportunity to reply, and to process and decide upon their complaints with priority. In addition, the state media authorities will also monitor compliance with Art. 20 EMFA, which, from 8 May 2027, requires manufacturers, developers and importers of devices and user interfaces controlling or managing access to and use of media services to enable users to change the settings, including the default settings, freely and easily, and to ensure that the visual identity of media service providers is clearly visible.

Art. 24 EMFA applies where a service operates an audience measurement system, for example where data on usage, reach, views, viewing time, interaction or target groups are collected and processed for decisions on advertising, pricing, purchases and sales, planning or distribution. Art. 24 EMFA requires audience measurement methodologies that are transparent, impartial, inclusive, proportionate, non-discriminatory, comparable and verifiable. Providers of proprietary audience measurement systems must, in addition, disclose their methodology, have it audited annually by an independent body and, upon request, provide media service providers with the audience measurement data relating to their offering, including non-aggregated data.

Another focus of the new Digital Interstate Media Treaty is transparency of ownership and shareholding structures in the media sector. The Commission on Concentration in the Media (KEK) is to maintain a publicly accessible and regularly updated database (Section 60(8) MStV-E). It is to contain, in particular, information on the owners and shareholdings of relevant media companies. This is intended to make it easier to understand who is behind a media offering and which economic interconnections exist. 

Political advertising under supervision

The draft allocates competences for the new EU Regulation on the transparency and targeting of political advertising. In future, the state media authorities are to monitor the obligations under Arts. 11 and 12 of the Regulation in respect of telemedia providers, and the obligations under Arts. 5, 7 to 17 and 21 in respect of media service providers and of telemedia providers that are not intermediary services within the meaning of the Digital Services Act (Section 111(4) MStV-E). This covers, in particular, requirements relating to the labelling, transparency and delivery of political advertising. Infringements are to be sanctioned under the federal Political Advertising Transparency Act (Politische-Werbung-Transparenz-Gesetz), which is still in the parliamentary process (Section 115(1b) MStV-E). 

AI-supported media supervision

Of particular practical relevance is the newly proposed express legal basis for the use of technical means in media supervision (Section 109a MStV-E). In future, the state media authorities may use automated systems that check text, image, audio and video content for possible infringements. This may also include content that is not freely accessible. The use of such systems is intended, among other things, to detect infringements of media law and youth protection in the media. 

At the same time, the draft contains safeguards: a possible infringement detected automatically must be reviewed by a human being without undue delay. If the suspicion is not confirmed, the personal data processed must be deleted; if no review takes place, they must be deleted after 30 days at the latest. Data required for supervisory proceedings after a suspicion has been confirmed are subject to strict purpose limitation and must be deleted after six months at the latest, unless supervisory proceedings are pending by then. In supervisory proceedings, the provider concerned must be informed that technical means were used. 

Online platforms may also be required to provide the state media authorities with suitable technical interfaces, insofar as this is technically and economically reasonable. The details are to be specified by the state media authorities in joint statutes; the main results of the use of technical means are to be reported annually. 

Competences at a glance

What the draft means for media services

Overall, the draft extends media regulation further into the digital space. Media houses, platform operators and advertising marketers have to prepare for new transparency, organisational and documentation obligations for their services, advertising processes and AI applications. In parallel, the states are deliberating on a second part of the Interstate Media Treaty. According to press reports, it is intended, among other things, to make providers of AI information systems responsible for the content they generate; no consultation draft has been published so far. It also remains unclear how such responsibility relates to the liability rules of the Digital Services Act where AI answers are integrated into search engines and platforms.