In this session of the AIDI Spotlights, we shifted our focus from Members organisation-wide work with AI and Data, and focused on a specific project. Deutsche Welle hosted the session and presented their work with AI4Media and Truly Media regarding Trustworthy AI for Content Verification: New AI-powered functions for the Truly Media tool and approaches to obtain more transparent AI components from external providers. 

Birgit Gray from DW first gave a brief overview of DW’s use case in the EU-co-funded project AI4Media. This entails a user-scenario-based introduction to the use of AI in content verification processes and an introduction to new AI-powered functions from external technology partners that have been integrated into a demonstrator of the Truly Media journalism tool. 

Birgit also summarised related challenges and learnings and explained further AI functions under development in AI4Media that are scheduled for integration. Following this overview, Birgit gave a demo of the new functions within the Truly Media tool. 

She then presented recent research in the field of “applied trustworthy AI”. This relates to approaches to obtain more transparent AI components from external providers, based on a cooperation with the provider of an AI-powered Deepfake-Detection-Function, CERTH-ITI. 

Finally, the session highlighted key results from AI4Media that are relevant for public service media providers in the EBU, and introduced related AI projects from DW.

Please contact Lalya Gaye for more information about how to attend this series of events if you are not an AIDI member.