When discussions turn to artificial intelligence in information warfare, the conversation quickly fixates on deepfakes and fabricated content. This focus obscures a more profound transformation: AI is fundamentally reshaping how information operations are designed, targeted and executed.
Beyond the obvious
"When people think about artificial intelligence in the information domain, they immediately think about content and deepfakes," observes Stéphane Grousseau, Deputy Director of Cyber Defence and Intelligence at Sopra Steria. "It's a genuine game changer, but content generation barely scratches the surface of the strategic implications."
The fixation of public and media attention on synthetic content is understandable. Yet this focus risks missing the essential point: AI tools are driving a paradigm shift in information gathering, analysis, targeting and strategic planning. And this could prove far-reaching in its consequences.
The massification revolution
Traditional disinformation campaigns suffered from a critical weakness: repetition. When hundreds of accounts shared identical messages, detection became straightforward.
"Artificial intelligence enables maintaining the same narrative whilst varying the wording and tone, which provides greater credibility," explains Grousseau. The result is not simply more content, but content that evades traditional detection methods. "One or two people saying something is suspicious. But 5,000 different sites telling the same story in different forms create an apparent consensus. Then you simply need to repeat the adage that there's no smoke without fire. And that's that. That is massification."
Agentic AI: strategic automation
For the expert, the latest technological development in artificial intelligence, agentic AI, marks a new turning point. "These are small AI agents, each with their own specialisation. This approach enables constructing extremely powerful information attacks with maximum automation."
Rather than a monolithic system, imagine specialised agents coordinating their actions: target identification, effective narrative analysis, personalised content generation, timing optimisation, real-time monitoring and adaptation. Each specialises in one task, and an orchestrator, itself an AI agent, coordinates the whole ensemble.
"We can create information attack templates, let them generate their content and react according to the situation, all at scales never previously envisaged," continues Grousseau. What required "bot farms with thousands of employees working for months" can now be accomplished with minimal personnel and resources.
Algorithmic profiling and micro-targeting
That's the attack manufacturing side. But today, AI tools no longer merely enable large-scale content generation: they can adapt to individual psychological profiles, exploiting the specific cognitive vulnerabilities of each target.
Algorithms analyse digital traces to construct detailed profiles of each individual's beliefs, fears and biases. Rather than flooding the information space with a generic message, AI systems can now deliver bespoke messages designed to maximise impact.
Multi-domain implications
AI's impact extends beyond cyberspace. Grousseau cites Ukraine's Operation Spider Web, a covert June 2025 drone attack that struck five Russian air bases deep inside Russia using 117 drones smuggled in trucks, causing an estimated $7 billion in damage: "We're facing a dual attack here.: "We're facing a dual attack here. The physical attack cost one lorry and around fifty drones. But the most powerful aspect is that it changed Muscovites' mindset by showing them that the war could impact their daily lives."
This synergy between kinetic action and information operation, amplified by AI, represents what the expert terms hybridity. Without the information dimension, the physical attack would have had limited impact. Combined with a sophisticated campaign, it maximised the desired psychological effect on the population.
Rethinking our defensive approach
AI tools impact far more than content generation. Consequently, all our defensive strategies must evolve. Sopra Steria is developing an operational approach that recognises this multi-dimensional reality. "We're creating a platform that integrates the entire chain: data collection, processing, modelling and capitalisation, analysis, decision support and response," explains Grousseau. This solution aggregates sovereign capabilities to detect, qualify and counter manipulation campaigns in real time.
Generative AI is transforming every aspect of information warfare: strategic planning, intelligence gathering, micro-targeting, large-scale automation, predictive analysis and real-time adaptation. Confronted with this systemic threat, the response must be equally comprehensive. Fortunately, through the Cercle Pégase, Sopra Steria is federating an ecosystem of technological and institutional actors to build robust defensive and response capabilities. Created in 2024 and coordinated by General Bruno Courtois, this think tank brings together over 400 experts from five complementary sectors (government, academia, media, business, and politics) to develop collective strategies against information manipulation and hybrid threats.