Disinformation has long been described as a parasitic 'noise' of the digital age. By 2025, it resembles more of an industry. The difference is not philosophical; it's operational: producing credible misleading content used to take time, a team, and the ability to create and disseminate narratives. Now, a handful of tools are enough to generate scripts, mimic a voice, create convincing visuals, adapt the same message into multiple languages, and test what 'performs' in real time.
The scale change is evident in global figures. Today, we are talking about approximately 5.24 billion active social identities worldwide, an immense field where speed takes precedence over verification. This mass creates a simple reality: the first content seen often becomes 'the true version,' even if the rebuttal follows later. The mechanism is known, but AI makes it more profitable. Where a content creator could publish a few videos per week, a semi-automated channel can publish several formats per day, optimize titles and thumbnails, and capture a stream of advertising revenue.
A recent case illustrates this shift: in the UK, an investigation uncovered networks of video channels disseminating false and sensationalist content, totaling nearly 1.2 billion views in 2025. This detail matters: we are no longer facing marginal rumors but volumes capable of influencing the public agenda, polarizing communities, and saturating attention. When the primary goal becomes monetization, the logic drives towards the most divisive, the most anxiety-inducing, the most 'shareable.' And when the goal becomes political, the same mechanisms are used to weaken trust, divide, and warp the debate.
European democracies have responded by putting pressure on platforms, not only in the name of ethics but also for political stability. The EU organizes 'stress tests' around elections to measure the ability of major platforms to limit the spread of manipulation campaigns, deepfakes, and coordinated networks. This movement reveals a truth: the digital space is now treated as critical infrastructure, on par with an electrical grid or a banking system.
Platforms find themselves oscillating between two contradictory imperatives. On one side, they promise anti-disinformation measures and labels against manipulated content. On the other, they depend on a model where engagement is the currency. Some innovations move towards traceability: Reddit, for example, is testing verified profiles in 2025 to limit impersonation and improve transparency. But the battle is asymmetrical: malicious actors innovate quickly, and AI reduces production costs faster than control systems improve their effectiveness.
In Morocco, the issue is not peripheral. The country has 35.3 million internet users at the beginning of 2025, and 21.3 million active identities on social media. This digital density has a mechanical effect: information is often created on platforms before stabilizing in media outlets. In this context, a well-staged rumor can travel faster than a verified report, and a deepfake can embed itself in the imagination before an official correction has time to deploy.
The real danger is not just believing a lie. It is being targeted. AI allows for creating messages tailored to specific segments: urban youth, diaspora, entrepreneurs, supporters, religious audiences, anxious publics, angry publics. Campaigns no longer need to be massive to be effective; they can be surgical. This is where disinformation becomes a weapon: it does not always seek to convince; sometimes it simply seeks to exhaust, to make everything dubious, to degrade the very idea of shared truth.
The response, however, is not to withdraw from the digital realm. It is to elevate collective reflexes. Content that triggers instant emotion, promises a 'revelation', cites 'anonymous sources', lacks a date, or exists only on a single platform deserves verification before sharing. Institutions must learn to respond quickly, with simple evidence, as slowness has become a disadvantage. Finally, the media have a role to play: in a saturated world, the value is no longer in publishing quickly, but in publishing accurately, and contextualizing with facts, figures, and names.
In 2025, the question is no longer 'Is there disinformation?'. It is: who has industrialized its production, with what means, and at what cost to public trust?
