AlgorithmWatch and partners urge German states to regulate how feeds and chatbots work in media treaty
Four civil society groups say the draft Digital Media State Treaty should drop accuracy percentages for AI answers and a one-in-five journalism quota, and focus on recommender design, user choice and oversight.
OpinionOpinion: the author's own view, summarised by us. It is not a news report or official guidance. By AlgorithmWatch, Agora Digitale Transformation, ISD Germany and Wikimedia Deutsch, civil society organisations responding to a consultation.

AlgorithmWatch, Agora Digitale Transformation, ISD Germany and Wikimedia Deutschland have submitted a joint statement on the discussion draft of Germany’s Digitale-Medien-Staatsvertrag (DMStV). With it, the federal states’ broadcasting commission plans to bring platforms and AI services under media law for the first time. The roughly 186-page draft of Part 2 covers AI information systems, artificially boosted reach, the findability of journalism, media concentration, advertising and supervision.
The groups argue that, instead of percentage figures for the ‘accuracy’ of AI answers, chatbots should give verifiable sources, flag uncertainty and face regular independent testing, especially before elections. They also criticise a quota requiring roughly every fifth post to come from a privileged pool, saying it changes what appears in feeds but not what they optimise for. In their view, users should have alternatives to feeds driven by personal data, and regulators should gain audit rights and researcher data access.
The comment period closed on 13 November 2026.
Editor’s note: This article relies solely on a statement published by the advocacy groups that submitted it.
Why it matters
If the draft keeps rules on AI information systems and recommender systems, platform and chatbot providers operating in Germany could face media-law duties alongside the DSA and the AI Act, including possible audit rights for state media authorities and data access for researchers. Teams that govern recommender or generative AI services should map which source-attribution, uncertainty-labelling and testing records they could produce. They should also track how the states reconcile the treaty with EU law, which the groups flag as a requirement.
Status: Consultation. Comment period on the Part 2 discussion draft closed on 13 November 2026
Sources
Was this useful?
Thanks. What was the problem?
Thank you — this helps us choose what to cover. If you spotted an error, please tell us what it is.
Prepared with AI assistance and reviewed by the editor before publication. Spotted an error? Tell us — see our editorial policy and methodology.
Governance areas
Who should care
- AI governance lead
- Legal & compliance
- Product owners for recommender systems