– VERification Assisted by Artificial Intelligence

Online disinformation and fake media content have become a serious threat to democracy, the economy, and society. Recent advances in AI now enable the creation of realistic synthetic content and its automatic dissemination through AI-driven bot networks. This makes it increasingly difficult for researchers and media professionals to assess the truthfulness and credibility of online content and recognize disinformation campaigns. aims to provide professional, trustworthy AI solutions against advanced disinformation techniques, developed in collaboration with media professionals and researchers, laying the foundation for future research in AI against disinformation. Core aspects of the activities are transparency (including explainability), robustness, and the ability to handle multilingual and multimodal content. The AI tools developed and used within are based on continuously collected fact-checking data derived from tens of thousands of pieces of "real-world" content. Furthermore, social media and web content can be analyzed and contextualized to detect disinformation campaigns and measure their impact. Results are validated by professional journalists and fact-checkers from the project partners (DW, AFP, EUDL, EBU) and associated partners. 

Responsibilities of Fraunhofer IDMT

Fraunhofer IDMT's main contributions to this project are in the following areas:

  • Research and Development on audio provenance analysis, to analyze the reuse and processing history of A/V content.
  • Research and Development on the classification of audio devices, audio scenes, acoustics events, etc., to support research and verification steps
  • Research and Development for audio manipulation and synthesis detection, with a special focus on context- and speaker-specific synthesis detection and dealing with combinations of so-called "shallow fakes" and "deep fakes"

Project Partners

Project Lead: Centre for Research and Technology Hellas (CERTH), Greece

  • The University of Sheffield (USFD), UK
  • Università di Urbino Carlo Bo (UNIURB), Italy
  • Fraunhofer Institute for Digital Media Technology IDMT, Germany
  • Universiteit van Amsterdam (UvA), The Netherlands
  • Kempelenov Institut Inteligentnych Technologii (KInIT), Slovakia
  • Università degli Studi di Napoli Federico II (UNINA), Italy
  • Centre Borelli, Ecole Normale Supérieure Paris-Saclay (ENS), France
  • Athens Technology Centre (ATC), Greece
  • Sirma AI EAD (trading as Ontotext) (ONTO), Bulgaria
  • Agence France-Presse (AFP), France
  • Deutsche Welle (DW), Germany
  • EU DisinfoLab (EUDL), Belgium
  • European Broadcasting Union (EBU) Switzerland

Funding is an EU project co-funded under the HORIZON-CL4-2021-HUMAN-01-27 (RIA) program.


September 2022 – August 2025

This might also be interesting for you

Research topic

Media Forensics

Trustworthy media content

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content

Research topic

Environmental Sound Analysis

AI-based analysis of complex acoustic scenes and sounds