AI's impact on fair competition in elections: an interview with cybersecurity expert Hicham El Amrani

While deepfakes are the most visible threat, they are only part of a much larger and more complex ecosystem. In an electoral or pre-electoral context, the threats posed by digital maliciousness multiply, and the stakes are much more serious. It can serve as a weapon of destabilization, public opinion manipulation, and even foreign interference.

As legislative elections approach, Morocco is primarily concerned. Recent examples of digital smear and disinformation campaigns remind us of the need for extreme vigilance. The first risk is audio and video deepfakes.

Fake voice recordings, now easy to produce by cloning voices from a few seconds of audio, are particularly concerning. More difficult to detect than doctored videos, they can spread false statements attributed to public figures and sow doubt in a few hours. But the most underestimated threat lies in large-scale automated disinformation.

Generative AI now makes it possible to produce thousands of credible fake articles, comments, testimonials, or profiles at a negligible cost. The goal is no longer just to spread a lie but to saturate the information space to make the truth indistinguishable from the false. Added to this are astroturfing, which artificially simulates a popular opinion movement through coordinated account networks, persuasive microtargeting, which customizes messages to influence or demobil