Digital sovereignty: the citizens’ blind spot

Europe wants “digital sovereignty” (“What is digital sovereignty?”, our interview with Julien Nocetti), but it continues to treat citizens as mere extras in a technological race between states and large companies. Behind the announcements of summits, investment plans, and “European preference,” a blind spot persists: the actual ability of each individual to understand, criticize, and master the technologies that are reshaping knowledge, work, rights, and democracy.

In the space of a few days, however, three political scenes painted the same picture. Berlin, France, and Germany have committed to accelerating the development of “cutting-edge” AI models, mobilizing talent and capital to catch up with the United States and China. In Brussels, the Commission presented a comprehensive digital simplification plan (digital omnibus), while proposing to water down the “AI mastery” requirement enshrined in the AI Act, reducing it to a simple encouragement left to the goodwill of companies. Finally, in Paris, the Assembly debated the fate of 4,000 digital advisors, whose funding remains uncertain, even though a significant portion of the population remains isolated from the opportunities offered by digital technology. Three moments, one paradox: Europe is investing heavily in AI infrastructure and models, but is devoting only crumbs to those who need to use them.

Linking AI, inclusion, and digital literacy

AI, inclusion, and literacy in digital technology, including AI, are still treated as separate issues, each with its own administrations, budgets, and timetables. Yet it is in their articulation that the famous European “third way” could emerge, between American hands-off and authoritarian control. Digital sovereignty would then be about more than just servers, data centers, and industrial champions; it would be about giving citizens real power to take action.

Today, the so-called digital divide translates into massive inequalities in access to rights, employment, public services, and information. In France, more than 30% of the population is considered digitally excluded. In social services offices, France Services centers, and neighborhood associations, the same difficulties arise: incomprehensible forms, hostile interfaces, and the requirement to “do everything online” without support. Nearly 65% of French people say they are afraid of digital tools, and 20% identify a lack of skills as the primary obstacle. At the European level, 44% of citizens lack basic digital skills.

This situation slows down the adoption of sovereign AI, which cannot be reduced to injunctions and communication campaigns inviting people to “dare” to use these devices.

Making skills a political choice

There are ways forward, but they are still treated as adjustment variables. First, we need to radically rethink the place of digital technology in schools, not just as a learning tool or a support for studying other subjects, but as a space for understanding the digitally-driven world in which children already live. This requires dedicated teaching hours, solid training for teachers, and institutional recognition of this knowledge on an equal footing with other subjects. This is what our collective supports in proposal 5 of its latest report on AI literacy, “Strengthening digital and AI education throughout initial training.”

Next, a portion of the billions of euros invested each year in AI could be allocated, by law, to skills development for all stakeholders (employees, trainers, local elected officials, community activists, citizens). This cannot be left to the goodwill of certain large companies through sponsorship or communication: it is a strategic investment to reduce inequalities, secure professional transitions, and strengthen critical analysis of uses. It is not just another tax, but a long-term investment that benefits both the private and public sectors. In France, the High Commission for Strategy estimated in 2018 that a targeted digital training plan for 4.7 million people would generate €1.6 billion in benefits per year. Yet we continue to move in the opposite direction.

Sovereignty is also at stake here.

Continuing to talk about “digital sovereignty” without setting out training and support obligations in black and white in European and national texts means accepting that this term only applies to governments and industry. Sovereignty reserved for experts is just a sophisticated form of expropriation.

This leads to an obvious contradiction: how can we expect millions of people to choose so-called sovereign infrastructures and tools if they do not understand the issues involved? As long as decisions are made “from above,” without massive investment in skills, they will remain abstract for the majority. Price, availability, efficiency, and security therefore remain decisive criteria, well ahead of any consideration of sovereignty. Digital and AI literacy must therefore be recognized as a democratic, social, and economic imperative, not as the last wheel of sovereignty.

It is time to consider each citizen not as a target for digital devices, but as an actor capable of questioning them and therefore choosing them in an informed manner. At the Berlin summit, Chancellor Friedrich Merz said that “digital sovereignty costs money, but digital dependence is even more expensive.” It will be even more if a whole section of society remains unable to acquire the skills needed to benefit from it.