epistemic sovereignty

A Manifesto

Whoever controls how you form knowledge controls what you think. This is the quiet law of our age, and almost no one is talking about it.

For three centuries, the sovereignty debate concerned territory, then data. We fought for borders, then for privacy. The next frontier is neither land nor information. It is cognition itself: the human capacity to construct knowledge through effort, doubt, memory, and judgment.

Artificial intelligence has ceased to be a tool and has become an epistemic infrastructure. It no longer merely answers our questions. Increasingly, it shapes which questions occur to us at all. When we delegate to machines not only the retrieval of answers but the very process of forming them, we transfer something deeper than data. We transfer cognitive autonomy.

The academic literature already has a name for this phenomenon: an epistemological rupture. The transition from embodied, effortful knowledge-making to instantaneous, machine-mediated cognition. A generation is now growing up inside interfaces that predict its thoughts before it has finished thinking them.

We are not against artificial intelligence. We build with it, work through it, and defend its transformative potential. Precisely for that reason we insist on a distinction that the current debate refuses to make: the question is not whether to use these systems. The question is whether we preserve the capacity to think without them.

Epistemic sovereignty is that capacity, held by individuals and by societies: the ability to form beliefs, reach conclusions, and exercise judgment through processes we understand, can inspect, and could, if necessary, perform ourselves.

This is not nostalgia. Defending memory, argumentation, problem-solving, and judgment is not a sentimental attachment to older ways of learning. It is a structural requirement of free societies. A democracy of citizens who cannot reason without an interface is a democracy in name only. A legal system whose practitioners cannot justify their conclusions independently of a model is a legal system running on borrowed authority.

The regulatory conversation, in Brussels, in Brasília, in Washington, remains fixed on risk, liability, and transparency. These matter. But they regulate the symptom. The cause is cognitive: the silent reallocation of thinking itself from humans to systems owned by a handful of companies. No liability regime addresses that.

We therefore hold:

• I.That cognitive autonomy is a foundational interest of the human person, prior to privacy and presupposed by every other right.
• II.That epistemic infrastructure is power, and concentrated epistemic infrastructure is concentrated power, deserving the same scrutiny we apply to concentrated economic or political power.
• III.That education, law, and public institutions have a duty to preserve spaces of cognitive effort: places where humans still reason, remember, argue, and decide.
• IV.That using artificial intelligence intensely and preserving epistemic sovereignty are not opposites. They are the two disciplines of the same freedom.

Epistemic sovereignty will be for the 2030s what data privacy was for the 2010s. The difference is that this time we can see it coming.

This site exists to start that conversation before it becomes urgent, because by then it will be too late to have it calmly.