News

Visit to the ECREA 2024 Our team members Edda Humprecht and Xin Zhou visited beautiful Ljubliana for this year's ECREA conference.

In his presentation “Leveraging OpenAI's multimodal APIs for analyzing digital propaganda in political communication”, Xin Zhou explained a new method for analyzing digital propaganda in social media using OpenAI's ChatGPT and Whisper APIs.

By converting video content into text and images, one can dive deep into political communication and identify themes, persuasion techniques and audience targeting. The pilot study with 30 political TikTok videos showed promising results, with the AI identifying the most important political topics with an accuracy of up to 85%!

ECREA was a fantastic opportunity to engage with experts and discuss how AI is revolutionizing the analysis of political communication. Thank you to everyone who attended!

Our team visits MISDOOM 2024 From September 2-4, 2024, Kostiantyn Yanchenko, Xin Zhou and Maria Grub attended MISDOOM in Münster.

The 6th Multidisciplinary International Symposium on Disinformation in Open Online Media attracted researchers from all over the world to discuss disinformation, propaganda, media literacy and much more, combining the disciplines of communication science and information systems.

Xin Zhou and Maria Grub took part in the PhD workshop before the conference. In small discussion groups, they presented their PhD projects on computational detection and analysis of multimodal propaganda content (X. Zhou) and the role of artificial intelligence in spreading, detecting and countering online disinformation (M. Grub) and received insightful feedback from top experts in the field. A big thank you to the organizers at the University of Münster and the mentors Thorsten Quandt, Kyoko Kuwahara, Jean-Christophe Boucher and Ofer Fridman for helping to move both projects forward.

Congratulations to Xin Zhou on his very first conference presentation, in which he presented the first study of his PhD thesis on analyzing implicit propaganda content on TikTok using a multimodal deep learning approach.