Towards Robustness of Disinformation Campaign Detection Algorithms in Open Online Media in the Context of Trustworthy AI
Overview
Open online media has become a ubiquitous source for information exchange, playing an important role in shaping public opinion and influencing individual perspectives. This influence, in combination with the open aspect of these media, is actively utilised to launch (dis)information campaigns, as demonstrated by preceding research around major events such as the German Federal election or the climate crisis. Detecting these organised information campaigns is essential to mitigate them, and TextClust, an online stream clustering algorithm, has demonstrated it can do this. However, it is important to adapt these algorithms due to the evolving nature of information attacks. To address this, we aim to combine state-of-the-art stream clustering algorithms with automated algorithm configuration, enabling robustness against new forms of (dis)information campaigns.
Key Facts
- Grant Number:
- 00.037.2024/ID 25306
- Research profile areas:
- Digital Humanities, Intelligente Technische Systeme
- Project type:
- Forschung
- Project duration:
- 05/2025 - 10/2026
- Funded by:
- Klaus Tschira Stiftung gGmbH
