Imen Chaanbi

Strategikia asks

With Mathieu Guidère, Imen Chaanbi co-wrote Géopolitique prédictive: l’intelligence au service de la prospective (Predictive Geopolitics: Intelligence in the Service of Foresight, Strategik.IA, 2025). The book combines artificial intelligence and strategic analysis to anticipate shifts across an arc stretching from the Maghreb and the Mashreq to the Sahel and the Gulf, and on to Afghanistan, Pakistan, India and China. Strategikia asked her about the method, its limits and what it changes for decision-makers.

1. How did this book, co-written with Mathieu Guidère, come about?

   From the meeting of two approaches. Mathieu Guidère founded predictive linguistics and has long worked on detecting signals in language. For my part, I support companies and institutions on the ground in Africa and the Middle East, with a very practical need: to anticipate rather than react.

   We wanted a book that would be useful both to decision-makers, who need clarity, and to analysts, who need operational tools.

2. Why did you choose this arc from the Maghreb to China?

   Because it is a space where tensions, transitions and power projection are closely intertwined. It has energy rivalries around the Gulf, shifting identities and the erosion of state mediation in the Sahel.

   It is a hard region to predict, and that is precisely what makes it valuable for foresight: it produces an abundance of the weak signals that herald tipping points.

3. In practice, how does artificial intelligence fit into your method?

   At three levels. First, a highly diverse database: economic series, climate data, migration flows, religious and political discourse, social media, videos and testimonies. Next, algorithmic processing that cross-references these data and brings correlations to light. Finally, and this is essential, human analysis that places those correlations within a strategic framework.

   The result takes the form of layered country profiles, covering security, economic, identity and environmental factors, regularly updated and designed to spot tipping points rather than describe a static situation.

4. Can you give an example of a weak signal?

   Southern Algeria, on the border with Mali and Niger, is a good example. On its own, no single indicator raises an alarm: advancing desertification, changes in certain religious discourse, the decline of social rituals, shifts in remittances. It is their combination that becomes significant.

   Likewise, in the Sahel, social unravelling was perceptible several months before the collapses became visible. That is what we call cascading conflicts: a local shock that spreads to neighbouring areas.

5. So can AI predict crises?

   No, and that needs to be said clearly. AI does not predict on its own: it makes emerging patterns readable. It provides trends and probabilities, never control over reality.

   Every prediction rests on choices of framing and variables, so it is never neutral. Hence the need for transparency: publishing the parameters, margins of error, sources and limits, and cross-checking the results against field expertise and local sources.

6. Is France lagging behind in this field?

   Foresight there too often remains theoretical and removed from operational needs, while countries such as Russia, China, the United Arab Emirates and India are investing heavily in these capabilities.

   Our book argues for a French culture of anticipation, one that accepts uncertainty as a given and builds on France’s own analytical traditions rather than importing foreign models.

7. You advocate “strategic optimism”. What is it based on?

   The classic Great Game, with stable players and clearly defined spheres of influence, has crumbled. Competition has become asymmetric, involving states, companies, non-state groups and platforms. The energy and technology transitions are opening up room for manoeuvre for players who have long been sidelined, particularly in the Global South.

   But anticipation has two faces. It can help us break out of reactive thinking, or it can fuel a race to predict and mutual suspicion. That is why we argue for shared anticipation, used to regulate rather than to justify unilateral decisions.