Artificial intelligence can prove to be an insidious weapon of mass disinformation. However, its impact remains limited, according to a report co-signed by EPFL scientists which analyzes the hundred or so elections taking place around the world in 2024.
From Taiwan to the United States via Senegal, AI has been at the heart of political campaigns, notes EPFL in a press release. Between “deepfake” and “chatbot”, the misuse of technologies destabilizes public confidence in the information disseminated by the media, whatever they may be, from television channels to social networks.
By analyzing the different campaigns and electoral results with local specialists, several experts note that AI has not actually changed the results of the elections. Its impact therefore remained limited. However, the proliferation of manipulated content, amplified by algorithms, has contributed to the fragmentation of opinions and fueled a climate of generalized distrust.
In Switzerland too
Deepfakes, audiovisual content artificially generated to imitate real people, marked the election year both in the United States and in Switzerland. In the report, the researchers recall that in view of the 2023 federal elections, Aargau SVP national councilor Andreas Glarner had created a deepfake video of his political rival Sibel Arslan.
Another example, cited in the report, notes that politicians like Donald Trump systematically describe information presenting them in an unfavorable light as “deep fakes”. While in India or Indonesia, the use of generative AI has crossed new boundaries, deceased political figures have been “resurrected” in the form of digital avatars to influence voters.
Increased capacities
The history of digital manipulation for propaganda purposes is not new. But AI has multiplied their capabilities. During election periods, the massive production and rapid dissemination of synthetic content – whether videos, images or texts – undermine public confidence.
In addition, the authors of the report observe a lack of regulation which has, consequently, allowed a proliferation of manipulative content, illustrating a legal void.
The analysis was carried out by the Initiative for media innovation (IMI) based at EPFL. It is published in the first issue of its magazine “Décryptage”.
Among the experts’ recommendations is the development of tools for traceability and detection of synthetic content. The importance of international regulation and accountability of technological players is also highlighted. Finally, the report highlights the crucial role of governments, businesses and civil society in creating an ethical and secure digital environment.
ats/vkiss
World
