Artificial intelligence has not merely added a new layer to foreign interference. It has altered its grammar. What was once slow, costly, and often clumsy has become scalable, adaptive, and disturbingly precise.
The report’s central insight is simple but uncomfortable. AI does not just amplify existing practices of influence and interference, it collapses the distinction between them. Influence becomes clandestine. Interference becomes normalized. The result is a continuous grey zone in which democratic societies struggle to identify when persuasion turns into manipulation, and when expression turns into attack.
From technological breakthrough to geopolitical instrument.
The report situates AI within a long historical arc, from early computational logic to contemporary generative models. Yet its real contribution lies elsewhere. AI is framed not as a neutral innovation but as a metatechnology, a system that generates systems, reshaping power relations across economic, military, and informational domains.
This matters because AI has become an explicit object of geopolitical competition. States no longer compete merely withAI but over AI. The race between the United States and China structures global technological hierarchies, while other actors adapt asymmetrically. Some invest in model development. Others specialize in strategic uses, particularly in information warfare. Russia, Iran, and North Korea are emblematic here. They may lag in large-scale innovation, but they excel at operationalizing AI for disruption, deniability, and cognitive pressure.
Hybrid warfare, scaled down to everyday life.
The report convincingly embeds AI-enabled interference within the broader logic of hybrid warfare. The key novelty is not violence but ambiguity. AI lowers the threshold of action. It allows hostile actors to operate below the radar of traditional legal and political responses, exploiting uncertainty, attribution problems, and institutional inertia.
Disinformation campaigns are no longer artisanal. Generative AI enables the mass production of plausible content, in multiple languages, tailored to specific audiences. Bots are no longer crude amplifiers. They simulate interaction, emotion, and controversy. Deepfakes are no longer exceptional events. They are part of a continuum that includes satire, parody, and “soft fakes”, eroding the very criteria by which truth claims are assessed.
The report’s detailed reconstruction of recent Russian operations targeting France illustrates this evolution with striking clarity. Campaigns such as Storm-1516 combined AI-generated content, automated dissemination, and strategic “laundering” through seemingly credible local websites. The objective was not persuasion in the classical sense. It was corrosion. Undermining trust, exhausting attention, and contaminating the information environment itself.
Democracies as preferred targets.
One of the report’s most sobering conclusions is that democratic systems are not merely vulnerable, they are structurally attractive targets. Openness, pluralism, electoral cycles, and media freedom create opportunities for manipulation that authoritarian systems do not reciprocally offer.
AI exacerbates this asymmetry. It is cheap to attack and expensive to defend. Producing disinformation at scale costs little. Detecting, verifying, and debunking it requires institutional capacity, human expertise, and time. The imbalance is not accidental. It is strategic.
The report also highlights a paradox that deserves more attention. While Europe has developed an advanced regulatory framework, notably with the AI Act and digital services regulation, its response remains fragmented. Legal sophistication does not automatically translate into operational resilience. Regulation governs platforms. Interference exploits ecosystems.
Regulating algorithms, protecting democracy.
The set of proposals advanced by the report is deliberately eclectic. They range from algorithmic transparency and labelling obligations, to pre-electoral safeguards, to the strengthening of specialized agencies such as Viginum and ANSSI. What unifies them is an implicit shift in perspective. Democracy is no longer protected solely through constitutional guarantees or electoral law. It requires informational security.
Some proposals are deliberately provocative. The idea of a pre-electoral “algorithmic reserve” raises difficult questions about freedom of expression and platform governance. So does the notion of an “artificiality score” attached to online content. Yet the report does not present these measures as silver bullets. It presents them as attempts to reintroduce friction into an environment designed for speed, virality, and emotional capture.
Equally important is the emphasis on societal resilience. Media literacy, journalistic standards, and public trust are treated not as soft values but as strategic assets. In that sense, the report quietly reframes counter-interference as a public good.
A legal and political dilemma, not a technical one.
What ultimately emerges from this analysis is a deeper dilemma. AI-driven interference cannot be solved by technology alone. Nor can it be neutralized through repression without betraying democratic commitments. The challenge is not to out-automate hostile actors, but to decide what kind of informational order democracies are willing to defend.
The report resists the temptation of technological determinism. It insists, instead, on political choice. Democracies must accept that vulnerability is part of their identity. But they must also refuse fatalism. Between naïveté and digital authoritarianism, there remains a narrow but vital space for democratic self-defence.
AI has changed the battlefield.
It has not changed what is at stake.