Month: January 2025

Abstract

The evolution brought by social media as well as the Internet itself has led to new paradigms in journalism and mass communication. New technologies revolutionize the way humans communicate and get informed about what is happening in the world; in parallel, however, individuals and organizations can exploit these new paradigms to pursue their own agenda. In this light, the spread of misinformation and disinformation can have a variety of negative effects on society, for instance, by creating dipoles in the political dialogue and threatening democracy, putting the health and security of citizens at risk through falsified information, or spreading conspiracy theories about climate change and the environment. To help in the fight against misinformation, the present study focuses on an innovative approach that evaluates and combines the results of different content verification services. This approach, entitled “Meta-Detection Toolset (MDT)” for content verification, consists of an algorithmic consensus (voting) mechanism based on weights, which rewards or penalizes verification services on the basis of their prediction results and the ground truth (verification labels) provided by human domain experts (fact checkers or other). Using a dedicated weight recalculation algorithm, as the feedback of the experts is gradually provided, the weights are recalculated and updated constantly, thus forming a continuous learning procedure for content verification.