On the evening of October 7, our fifth MediaTech Hub Turntable took place at the UFA premises. This time, the focus was on the challenges and solutions for detecting and combating deepfakes in editorial processes.
Tim Polzehl, winner of our first MediaTech Hub Demo Day and CEO of Gretchen AI, provided the technical introduction. He offered exciting insights into his company’s practical experience in developing AI-supported systems for detecting manipulated audio, image, and video content.
In a focused discussion, participants talked about how editorial teams can set up their tech infrastructure, internal processes, and staff skills so they can spot deepfake manipulations early on.
The main takeaways from the evening were:
- Technical infrastructure: Editorial teams need closely integrated AI tools, clearly defined detection processes, and regularly trained staff in order to reliably detect attempts at manipulation.
- Trust management: After a deepfake incident, transparent communication is essential. Only through openness and ongoing media literacy work can lost trust be restored.
- Strategic partnerships: Collaboration with external fact-checking organizations and AI startups promotes the further development of technical, methodological, and ethical standards.
We would like to thank all participants for their enthusiastic dialogue and look forward to exploring further perspectives and best practices in the upcoming Turntable.
