Fourteen teams, two days, one task: During the inaugural Hackathon by the AI for Media Network, participants faced the challenge of swiftly turning their innovative AI ideas for journalism into practical applications. A team ultimately emerged victorious by enhancing story depth with the help of Large Language Models.
Held from July 15 to 17 at Microsoft’s Munich branch, the Hackathon focused on “Generative AI and Automated Workflows in Media Production.” Organized by BR, Public Value Technologies —the digital subsidiary of BR and SWR—and Microsoft, the goal was to create AI-driven applications to enhance efficiency in the media sector.
A total of 69 individuals formed 14 interdisciplinary teams composed of developers, data analysts, content creators, project managers, and UX designers.
Hackathon Format and Resources
Over two days, participants developed their prototypes, utilizing Microsoft’s AI services, such as transcription tools,text and image generators. On the third day, each team had five minutes to present their results and convince both the audience and a nine-member jury comprising:
- Lilian Dammann, Director of Analytics & Data Platforms, pub.
- Hans Hoffmann, Head of Media Fundamentals and Production Technology, European Broadcasting Union
- Carmen Heger, Director Data & AI, Süddeutsche Zeitung Digital Media
- Jan Ippen, Managing Director, Ippen Digital
- Gert Kauntz, Managing Director, pub.
- Uli Köppen, Chief AI Officer, Bayerischer Rundfunk
- Axel Kress, Azure Core Sales Lead Telco, Media & Professional Services – SAPonAzure Sales Lead, Microsoft Deutschland GmbH
- Patrick Markowski, Head of Editorial Tech & Operations, BILD
- Christian Nitsche, Editor-in-Chief, Bayerischer Rundfunk
Winning Team and Application
The jury evaluated the teams based on content relevance, technical complexity, user experience, and product maturity. Team “Accio” won by creating an application that enriches news with relevant archive information. Named after a summoning spell from the Harry Potter series, “Accio” supplements news reports with background information, automatically generating and answering questions with data from archives. This allows for the rapid creation of comprehensive articles from straightforward news reports.
Runner-Up Projects
Second place went to an ARD team that developed “Study Buddy,” an application that makes ARD educational content accessible using AI. It features scenarios for preparing presentations and exams. The third-placed team, “Verifuchs” from Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS, created a verification tool to assess the trustworthiness of text using a simple traffic light scoring system.
Future Steps
Determining which projects will move forward is the next step. Interested parties can contact aiformedia@br.de for more information. To experience the Hackathon from a participant’s perspective, watch our recap video.
About AI for Media Network
The AI for Media Network is an initiative by Bayerischer Rundfunk and Ippen Digital, designed to be a practical, interdisciplinary platform where members can showcase and develop AI solutions for editorial use cases. It connects public broadcasting corporations, media houses, industry, and academia, exemplified by this Hackathon on generative AI in media.
