Artificial intelligence is reshaping the strategies behind information manipulation, particularly in electoral contexts, and Morocco faces this risk directly as legislative elections approach, Le Matin daily reports. While deepfakes remain the most visible threat, generative AI now allows deceptive content to be produced and distributed at unprecedented speed and scale, with experts identifying five main categories of risk alongside a broader detection gap.
The first is audio and video deepfakes, which can show a candidate appearing to say or do things that never happened. Such content is effective mainly because of its emotional, seemingly concrete nature, which can spread and shape public conversation before verification takes place; experts note voice cloning is increasingly accessible and may pose an even greater risk than video, since a convincing fake call or voice message can exploit the trust people place in what they hear.
The second is automated disinformation at scale, where generative AI can rapidly produce misleading articles, comments and fabricated online identities, saturating the information space and making it harder for citizens to separate verified information from a flood of competing narratives. The third is astroturfing, coordinated networks of accounts that simulate grassroots consensus, artificially amplifying a topic or attacking a target while appearing to reflect organic public sentiment.
The fourth risk is persuasive micro-targeting, where AI combines behavioral data with content generation to tailor messages to narrow voter segments, which can be used not only to persuade but to demobilize specific groups by suggesting that voting is pointless. The fifth is foreign digital interference, where state or state-linked actors could combine these techniques with hacking and coordinated leaks to undermine trust in institutions rather than favor a specific outcome.
Beyond these five threats, experts point to a deeper vulnerability: a detection blind spot, since the most effective operations often stay subtle enough to avoid early attention. Le Matin daily argues that countering AI-assisted manipulation requires moving beyond fact-checking toward sustained monitoring of how content spreads, which accounts amplify it, and what coordination patterns underlie its diffusion, treating vigilance as a continuous capability rather than a one-off reaction.
