AI poses five major threats to election integrity in 2026

While deepfakes are the most visible threat, they are only part of a larger ecosystem. Generative AI can produce deceptive content, create fake identities, automate their dissemination, personalize messages, and coordinate complex influence operations at unprecedented speed and scale. This shift means manipulation is no longer just about falsified content but can become a factory for manipulation and influence.

In electoral settings, disinformation can harm a candidate's reputation, influence perception of an issue, exacerbate divisions, discourage participation, or foster distrust in institutions. In severe cases, it can also be a tool for foreign interference. Morocco, with upcoming legislative elections, is particularly concerned, as digital smear and disinformation campaigns have shown the threat is no longer theoretical.

Five main threats are identified, along with a sixth transversal risk called the "detection blind spot." The first threat is audio and video deepfakes, which can create fabricated sequences of candidates saying or doing things they never did, causing significant damage before authenticity can be verified. The second threat is large-scale automated disinformation, where generative AI can rapidly produce articles, comments, social media posts, fake testimonies, or responses to artificially fuel controversies, creating credible digital