Several State media authorities in Germany are, according to corroborating MDRreports dated 5 November 2026, using artificial intelligence to automatically screen public posts on social networks and websites. The system named KIVI flags suspected hate speech, banned Nazi symbols and content relevant to the protection of minors. According to the reports, authority staff then make the decisions; the software does not take any final action.
Supporters point to the sheer volume of publicly accessible content. Automated preliminary screening is intended to help identify criminal posts more quickly and, where necessary, pass them on to Polizei or Public Prosecutor’s Office . The media regulator also argues that large platforms such as YouTube and TikTok must cooperate technically so that searches can be carried out lawfully and efficiently.
Criticism has come, among others, from journalists’ associations and opposition politicians in Saxony. They warn of indiscriminate, blanket scans and see risks to the right to informational self-determination. They are calling for strict thresholds for intervention, a clear restriction of purpose, independent oversight and transparent procedures for reviewing, documenting and deleting matches. The monitoring of messaging services such as Telegramis also viewed with scepticism, as it could easily lead to large collections of data whose relevance and proportionality are disputed.
The legal position has not yet been conclusively settled. Proposals to define powers for automated analysis precisely are under discussion at federal and state level. In particular, reform of the Interstate Media Treaty is being considered to set out responsibilities, transparency obligations, and data retention and deletion periods. Independently of this, criminal-procedure measures remain subject to the existing legal thresholds; there is currently no comprehensive legislative solution.
The debate brings together practical necessities and fundamental rights issues. The facts are that screening concerns publicly accessible content, matches are verified manually, and, according to the media authorities, there is no direct state mass surveillance. Attributed positions: critics describe a system that interferes with freedom of expression without any specific suspicion. Predictions: whether future rules will strike a balance between effectively protecting those affected and strictly controlling interventions will be decided in the ongoing parliamentary discussions.
It remains unclear how detailed the obligations on platforms to cooperate and provide interfaces will be, and what independent oversight will regularly review the procedures. One thing is clear, however: without transparent search criteria, documented handovers, and short retention and deletion periods, a lasting and legally robust practice is unlikely to endure.
Editor: Mira Cadena (Article created with AI assistance).