• Priority Area Supporting Research into Media, Disinformation and Information Literacy Across Europe
• Year 2023, 2024, 2025
• Country Greece
• Project Status Ongoing

The task of fact-checking is often laborious and time-consuming, struggling to keep pace with the rapidly growing volume of online misinformation. The emergence of new social networks, the increasing use of information modalities other than text, and the use of opaque recommendation algorithms make the fact-checking process all the more challenging. Given this situation, it is critical to develop new tools that provide accurate claim detection and matching solutions that can accelerate and apply fact-checking processes in a range of contexts.

The MuseAI project focuses on developing a multilingual claim detection and matching solution. In contrast to the majority of research in this area, mostly focused on English or monolingual approaches, the proposal adopts an incremental approach, starting with Spanish and English and gradually incorporating more European languages, through the use of NLP multilingual-based machine learning models, thus tackling the problem from its global perspective.

Project’s approach has several novel aspects: 1) It blends advanced natural language processing and machine learning techniques for increased accuracy and efficiency, analysing context and language semantics; 2) It enables us to understand misinformation dynamics on emerging platforms like TikTok, as well as AI’s role in the generation of content, and how recommender systems work; 3) It creates open-source software and a new misinformation detection framework for fact-checkers. The MuseAI solution will help fact-checkers scale up their tasks associated with monitoring the web to detect fact-checkable claims, as well as repetitions of a known false claim. It will also support other professionals studying misinformation like researchers, analysts, and investigators to have a better understanding of the disinformation dynamics in play.