Fighting unlawful content on the internet with AI-based monitoring

Presenter(s): Jan Thomsen (Condat AG)

Berlin-based Condat AG has developed an AI-based monitoring solution used by German state media regulation authorities to detect unlawful content on the internet. The solution consists of an automatic search component for the most common social media platforms, an automatic analysis and classification component which employs a variety of different AI/ML services for the automatic detection of potential law infringements, and a UI/UX and workflow component for the manual evaluation of the results the platform has generated., including generating reports for legal complaints. This talk describes the platform, the different AI/ML methods which are used, and the detection success rates with respect to different infringement categories.