Research

In a global context marked by political polarization and the proliferation of fake news, artificial intelligence (AI) has become an essential tool to combat disinformation. This is explained in the chapter written by Roger Cuartielles (Pompeu Fabra University) and Uxía Carral (Carlos III University of Madrid), both collaborators of Learn to Check, ‘Artificial intelligence tools against populist disinformation’ in the book ‘Democracy at risk: Internet and AI at the service of populism?’ (ed. UOC), which analyzes how AI is transforming strategies for verifying and identifying misleading news, especially during electoral periods.

According to the authors, disinformation has become an increasingly common phenomenon, driven by the strategic use of populist narratives that discredit traditional media and present institutions as corrupt. This type of discourse fosters an environment in which fake news spreads more easily, especially through social networks and instant messaging applications.

During election campaigns, this practice intensifies, as political groups can use disinformation to mobilize voters or discredit their opponents. The authors of the chapter point out that this dynamic not only affects public debate, but also erodes trust in the democratic system and traditional media.

This work highlights that, in this scenario, AI-based tools have been developed to combat the spread of fake news. One of the pioneering initiatives in this field is ClaimBuster, a system created by the University of Texas that uses machine learning and natural language processing to identify political claims with potential for verification. This system was first applied during the 2016 US election debates and has since evolved to offer a more precise analysis of political information.

Several automated verification initiatives have also been implemented in Europe and Latin America. The Argentine platform Chequeado developed Chequeabot, a system capable of transcribing videos and identifying claims that can be verified. In the United Kingdom, Full Fact has implemented tools such as Live and Trends, which allow real-time monitoring of the spread of disinformation and tracing the path of a fake news story as it spreads.

The chapter focuses on the Spanish general elections of July 23, 2023, where artificial intelligence was used significantly to combat disinformation. Platforms such as Newtral, Maldita.es and EFE Verifica developed automated systems to filter and analyze political information.

Despite the advances, the text notes that the implementation of AI in fact-checking is still in its early stages. Most current tools focus on the early stages of the process, such as the detection of potentially false claims (claim detection) and their identification (claim spotting), but there is still a long way to go in aspects such as final verification and the creation of detailed reports (claim matching, claim checking and fact-check reporting).

The authors conclude that, to maximize the effectiveness of AI in the fight against disinformation, it is essential to complement the technology with human oversight. They also propose closer collaboration between verification platforms, media and academic institutions to develop more sophisticated and efficient tools.

With the exponential growth of AI and its impact on digital communication, the challenge for fact-checkers and researchers is to ensure that these technologies are used ethically and transparently, avoiding algorithmic biases and improving the ability to respond to disinformation.

More information about this publication here.