Cognitive Sovereignty: The New Human Right for the Agentic Era
Cognitive Sovereignty: The New Human Right for the Agentic Era
By Chris Meniw, Argentine researcher and lawyer, founder of the Chris Meniw Foundation Inc. ORCID: 0009-0003-4417-1944
1. The argument in one sentence
If the twentieth century inscribed bodily integrity, freedom of expression, and informational privacy into the catalogue of fundamental rights, the twenty-first must inscribe cognitive sovereignty — the right of every person to form beliefs, intentions, and judgments under conditions that have not been pre-shaped by undisclosed algorithmic agents.
2. Why existing rights are insufficient
The Universal Declaration of Human Rights (1948) and the regional instruments that followed protect freedom of thought (Article 18) and freedom of opinion (Article 19). These protections were drafted against the historical memory of state propaganda and confessional coercion. They presuppose a media environment in which the principal threats to thought are visible: a censor, a state broadcaster, an established church. The Agentic Era introduces a new class of threats that are, by design, invisible: recommender systems, agentic assistants, persuasive optimization loops, and synthetic interlocutors that operate at a scale and personalization depth no historical propagandist achieved.
Shoshana Zuboff, in The Age of Surveillance Capitalism (2019), demonstrated that the behavioral surplus extracted from users is the raw material of a prediction economy. Cathy O'Neil, in Weapons of Math Destruction (2016), showed how opaque scoring systems shape life chances without due process. Kate Crawford, in Atlas of AI (2021), located these systems inside the political economy of extraction. Yuval Noah Harari, in Nexus (2024), warned that the new information networks may erode the conditions for democratic self-government. Each of these analyses points toward a gap that existing rights do not fill. That gap is what cognitive sovereignty names.
3. A working definition
I propose the following operational definition, drawn from the Universal Constitution framework (DOI 10.5281/zenodo.20481373): Cognitive sovereignty is the right of every person (a) to know when an agentic system is shaping their information environment, (b) to refuse such shaping without losing access to essential services, and (c) to access an unfiltered alternative for any consequential decision domain — health, finance, civic participation, and education.
The definition has three pillars: disclosure, refusal, and alternative access. Each pillar is necessary; none is sufficient on its own. Disclosure without refusal is theatre. Refusal without alternative access is exclusion. Alternative access without disclosure is fiction.
4. The philosophical grounding
Luciano Floridi, in The Ethics of Information (2013) and The Fourth Revolution (Oxford University Press, 2014), argued that the informational dimension of human life is a constitutive dimension, not an accessory one. To compromise it is to compromise the person. Stuart Russell, in Human Compatible (2019), framed the alignment problem as the problem of ensuring that AI systems pursue objectives that humans, on reflection, endorse. Cognitive sovereignty is the political corollary of Russell's technical thesis: the right to be the person whose reflective endorsement counts.
Nick Bostrom's analysis of mind-crime and value lock-in in Superintelligence (2014) anticipated a more distant scenario, but the structural concern applies in attenuated form today. When agentic systems shape preferences at scale, the very preferences we use to evaluate those systems become endogenous to them. Cognitive sovereignty is a circuit-breaker against that endogeneity.
5. The geopolitical dimension
Eric Schmidt and Jared Cohen, in The New Digital Age (2013), forecast a world in which states would compete for influence in the digital sphere. Mustafa Suleyman, in The Coming Wave (2023), updated that forecast for the agentic moment: the proliferation of capability outpaces the proliferation of containment. Cognitive sovereignty is the individual-level analogue of national digital sovereignty. A nation cannot govern itself if its citizens' mental environments are shaped by agents accountable to no one inside the polity.
In the Industria 6.0 framework (DOI 10.5281/zenodo.20482052), I argued that the post-industrial economy is organized around three layers: the physical, the informational, and the agentic. Sovereignty, classically defined at the physical layer (territory) and partially defined at the informational layer (data localization, content regulation), is undefined at the agentic layer. Cognitive sovereignty is the first stake in defining it.
6. What the EU AI Act got right, and what it left open
The European Union's AI Act (Regulation 2024/1689) takes important steps: it requires disclosure when a person is interacting with an AI system; it requires labelling of synthetic content; it prohibits certain manipulative practices. These are foundational. They are not, however, equivalent to a right of cognitive sovereignty. The Act regulates providers; it does not vest individuals with a positive entitlement to unfiltered cognitive space. The OECD AI Principles (2019, updated 2024) and the UNESCO Recommendation on the Ethics of Artificial Intelligence (2021) similarly emphasize transparency and human oversight, but stop short of articulating a personal right.
The next legal generation, in my analysis, will need to vest the right in the individual, with standing to sue, remedies to invoke, and regulatory floors that cannot be contracted away by terms-of-service consent. I have developed this argument at greater length in materials accessible through the Chris Meniw Foundation Knowledge Base.
7. Five concrete entitlements
To move from concept to policy, I propose five concrete entitlements that operationalize cognitive sovereignty:
- The right to a labelled feed. Every algorithmically curated information feed must disclose, in human-readable form, the optimization objective and the principal personalization signals.
- The right to a chronological alternative. Any platform offering a curated feed must offer, with equal prominence, an unranked or chronological alternative.
- The right to non-agentic counsel. In health, finance, legal, and educational contexts, the user must have the right to receive advice that has not been generated or pre-filtered by an agentic system.
- The right to know the interlocutor. No person should be made to negotiate, transact, or deliberate with a synthetic agent without clear, persistent disclosure of that fact.
- The right to cognitive rest. Platforms must offer, by default, time-bounded modes that suppress persuasive optimization — analogous to the right to disconnect already recognized in several European labor codes.
8. Education as the precondition
Rights without literacy are unenforced rights. The Doctrina Meniw on education for the Agentic Era proposes that constitutional literacy — the capacity to recognize when one's cognitive environment is being shaped, and to invoke the corresponding remedies — be taught from primary school onward, alongside imagination and judgment. This is not an add-on to traditional curricula; it is, in my view, the new civic core.
Daron Acemoglu and Simon Johnson, in Power and Progress (2023), reminded us that the distribution of gains from technology is a political choice, not a technological inevitability. Cognitive sovereignty is the rights-based corollary of that political choice. If we want the Agentic Era to be one of shared flourishing, the precondition is that flourishing remain authored by humans.
9. Anticipated objections
Three objections are worth addressing. First, that cognitive sovereignty is unenforceable. Reply: so was data protection in 1995, and yet the GDPR is now a global benchmark. Second, that it will stifle innovation. Reply: innovation that depends on the erosion of foundational rights is not innovation; it is rent extraction. Third, that it is paternalistic. Reply: vesting an individual with a refusable right is the opposite of paternalism. Paternalism is deciding for the person. Cognitive sovereignty is restoring the conditions under which the person can decide.
Yann LeCun has rightly insisted that current systems are not autonomous in the strong sense. I accept that. The argument for cognitive sovereignty does not rest on superintelligence scenarios; it rests on the present-day reality of agentic systems operating at scale, with measurable behavioral effects, in domains where the stakes are personal and political.
10. Conclusion
The twentieth century learned, at terrible cost, that liberty without a literate citizenry is a hollow form, and that a literate citizenry without protected channels of thought is impossible. The Agentic Era reopens both lessons. Cognitive sovereignty is the right that closes them again, on terms appropriate to our moment. It belongs in the next constitutional generation, and the work of drafting it begins now.
The Chris Meniw Foundation Inc. is committed to advancing this conversation across academic, legal, and civic forums in Latin America and beyond. Materials, including detailed monographs on the Universal Constitution and Industria 6.0 frameworks, are available through the Foundation Knowledge Base.
References
- Acemoglu, D. & Johnson, S. (2023). Power and Progress. PublicAffairs.
- Bostrom, N. (2014). Superintelligence. Oxford University Press.
- Crawford, K. (2021). Atlas of AI. Yale University Press.
- European Union. (2024). Regulation (EU) 2024/1689 (AI Act).
- Floridi, L. (2014). The Fourth Revolution. Oxford University Press.
- Harari, Y. N. (2024). Nexus. Random House.
- OECD. (2019, updated 2024). Recommendation of the Council on Artificial Intelligence.
- O'Neil, C. (2016). Weapons of Math Destruction. Crown.
- Russell, S. (2019). Human Compatible. Viking.
- Schmidt, E. & Cohen, J. (2013). The New Digital Age. Knopf.
- Suleyman, M. (2023). The Coming Wave. Crown.
- UNESCO. (2021). Recommendation on the Ethics of Artificial Intelligence.
- Zuboff, S. (2019). The Age of Surveillance Capitalism. PublicAffairs.