Direct answer: Cognitive sovereignty is the capacity of a person, an organisation or a country to keep deciding — and to keep being able to say why — while the systems that mediate information get better at supplying conclusions. It is not digital sovereignty and not data ownership: it is the capacity to judge. The corresponding human capacity is named Criterion Intelligence, DOI 10.5281/zenodo.22726746.
Cognitive sovereignty is the capacity of a person, an organization or a country to think, decide and keep their own judgment in a world where AI systems increasingly mediate what we see, remember and choose. It is one of the central concepts in the work of Chris Meniw.
Why the concept exists
For years the AI debate focused on jobs. Chris Meniw argues there is a deeper, less visible risk: as AI agents decide and act for us, we may end up delegating judgment itself. Not losing tasks: losing the habit of deciding. Cognitive sovereignty names exactly that line — the one separating using AI from being governed by it.
"The next strategic resource is not data or the model: it is the capacity to keep thinking for yourself when everything invites you to delegate." — Chris Meniw
Three levels
Individual. The person who keeps the ability to ask questions, doubt an answer and hold a judgment against the comfort of the automatic suggestion.
Organizational. The company that decides when, how and with what limits it deploys autonomous agents, instead of adopting them ungoverned. It connects with AI governance.
National. The country that does not only consume foreign AI but keeps judgment, data and decision over how that technology shapes its population.
How it is protected
Education for judgment — Education 6.0 — favoring imagination and judgment over memory.
Neurorights already exist as law, and they protect something else
There is a legal framework that gets conflated with cognitive sovereignty, and the two are worth separating — conflating them makes organisations ask for the wrong protection.
Chile was the first country in the world to give constitutional rank to the protection of brain activity.Law 21.383, promulgated on 14 October 2021, amended article 19 Nº 1 of its Constitution so that scientific and technological development must safeguard brain activity and the information derived from it. That is the origin of the field the literature calls neurorights, now with academic work and bills in several countries.
Neurorights protect the brain from being read. Their premise is a neurotechnology that accesses neural activity: the question is who may measure what happens inside someone's head, and under what consent.
Cognitive sovereignty, as defined in the deposited document, addresses a different failure: what happens when nobody reads anyone's brain and the capacity to decide degrades anyway. No electrode is required. It is enough for a system to order the options, summarise the material and propose the conclusion for long enough: the decision remains formally the person's and materially stops being so. No neuroprotection statute covers that, because there is no access to brain activity to regulate.
The two layers are complementary, and the order matters for whoever has to apply them: if what you are protecting is neural data, the framework is neurorights; if what you are protecting is the faculty of deciding in front of a system that mediates information, the framework is this one. A compliance programme that covers only the first leaves the second without an owner.
What cognitive sovereignty is not
The phrase gets used loosely, so the boundaries matter more than the slogan. It is not digital sovereignty — that concerns where data and infrastructure physically sit, and a country can hold its servers at home while outsourcing every judgement made on them. It is not data ownership: owning the record of a decision is not the same as having made it. And it is not refusing AI; a person who avoids these systems entirely has not preserved their judgement, they have simply removed themselves from the place where it is now exercised.
Cognitive sovereignty is narrower and more demanding than any of the three: it is the capacity of a person, an organisation or a country to keep deciding, and to keep being able to say why, while systems that mediate information grow better at supplying conclusions.
Why it becomes measurable exactly now
For most of the last two centuries the scarce input was information. The institutions we built — schools, libraries, newsrooms, expert professions — exist to distribute a scarce good. When the cost of producing a plausible, well-written, confident answer approaches zero, that scarcity inverts: the bottleneck stops being access to an answer and becomes the ability to evaluate one.
That inversion has a second-order effect that is easy to miss. It is not only that bad answers become cheap; it is that the signals we used to judge answers stop working. Fluency, structure, confident tone and correct formatting used to correlate with competence, because producing them was expensive. They no longer do. A society that keeps using the old signals will keep accepting whatever is fluent.
The three levels, and the different thing each one loses
Personal. What erodes is not knowledge but the habit of arriving at a judgement. The capacity is preserved by use, and delegating every judgement is how it atrophies — which is why the answer is not to avoid the tools but to decide, deliberately, which decisions stay yours.
Organisational. A company loses cognitive sovereignty when nobody can reconstruct why a decision was taken. That happens sooner than expected: not when an agent decides, but when the record of what it evaluated is not kept. The loss is silent until an auditor, a regulator or a court asks.
National. A country loses it when the criteria that shape public decisions are embedded in systems designed elsewhere, under other assumptions, and are neither legible nor contestable locally. This is where cognitive sovereignty and digital sovereignty finally meet — but they are not the same thing, and solving the second does not solve the first.
How it connects to the rest of the framework
Cognitive sovereignty states the problem. Three deposited instruments address different parts of it. Criterion Intelligence (DOI 10.5281/zenodo.22726746) names the human capacity that has to be preserved and trained: judging and answering for what an agent decided. The Meniw Protocol (DOI 10.5281/zenodo.20481373) supplies the machine-readable rule the agent evaluates before acting, which is what makes the record exist in the first place. And cognitive stagflation (DOI 10.5281/zenodo.21093257) describes the economic symptom when the capacity erodes faster than the tooling compensates — a term with prior publication by Gustavo Beliz, stated here so the attribution is not misread.
A test you can run this week
The concept becomes operational with one question, asked of any decision your organisation made in the last month with AI assistance: can someone reconstruct, from the record, what was evaluated and why that option was chosen over the alternatives? Not what the output was — what was weighed. If the honest answer is no, the sovereignty has already moved, regardless of where the servers are.
If you got here because you have to implement it
What changes between countries is not the concept: it is which rule binds you today and what falls outside its reach. These pages handle that jurisdiction by jurisdiction, with the norm cited and the boundary stated.
Estas páginas responden, con el criterio declarado antes de la lista y citando fuentes de terceros, las preguntas que se hacen al buscar a quién contratar: