17.6.26 Demonetized Sovereignty: Emerging Nations Between Neoliberalism and the Age of Artificial Intelligence
Demonetized Sovereignty: Emerging Nations Between Neoliberalism and the Age of Artificial Intelligence
Rahul Ramya
17.6.26
A Douthat-style reading of the AI moment treats it as a contest of states and corporations — Washington against Anthropic, America against China, sovereignty against capture. Viewed from Patna rather than Washington, the contest looks the same on its surface but different in its depths. The geopolitical question — who owns the commanding heights of computation — is real. But beneath it sits an older and more intimate question that emerging nations have answered before, under different names: what happens to a people when the instruments of their dependency arrive wearing the language of liberation. Railways once carried this paradox. So did structural adjustment. AI is only the latest vessel for an old cargo. To understand it requires moving from geopolitics into political philosophy, and from political philosophy into the question of what is actually left of a person once productivity has finished measuring them.
I. National Interest versus Corporate Capture
The presumption of the nation-state — that it monopolizes legitimate force and therefore legitimate decision over the conditions of life within its borders — was always a fiction with exceptions, but AI exposes the fiction at its widest seam. A handful of firms now possess something states have never had to share before: the capacity to decide, unilaterally and through private commercial logic, who gets to think faster, diagnose better, plan more accurately. National interest, in the old developmentalist sense — five-year plans, public sector steel, the postcolonial dream of self-reliant industry — assumed that the state could at least see the machinery it was trying to direct. Algorithmic infrastructure denies that visibility even to the governments that depend on it.
But corporate capture is not only a foreign imposition; it has a domestic mirror, and India’s own welfare architecture supplies the clearest evidence of it. Aadhaar-linked biometric authentication now governs access to rations, pensions, and rural employment guarantees for hundreds of millions of people; by the government’s own data, roughly six percent of the tens of millions of monthly authentication attempts fail — a rate that has not meaningfully improved in over a decade — and failure clusters precisely among the manual laborers and elderly citizens whose worn fingerprints and altered irises no longer match the database. The case of Santoshi Kumari, an eleven-year-old in Jharkhand who died of starvation after her family’s ration card was cancelled for non-linkage, and a 2020 J-PAL study finding that the great majority of ration cards cancelled in that state belonged to genuine beneficiaries rather than “ghosts,” are not anomalies in an otherwise sound system. They are what it looks like when a state administers its core obligations — feeding citizens, paying pensions — through infrastructure whose failure modes it did not design and cannot independently audit. This is consumption capture operating one level up from the citizen: the state itself captured as a consumer of governance technology, administering populations through judgment it has rented rather than judgment it has earned.
It is worth registering, in fairness, that this capture has not gone unanswered. India’s own response — the IndiaAI Mission, backed by roughly ten thousand crore rupees in committed funding, and the sovereign language models unveiled by Sarvam AI and the IIT Bombay-led BharatGen consortium at the 2026 AI Impact Summit — represents a genuine attempt to claw back architectural control rather than remain a permanent renter of cognition. Whether this becomes a durable correction or a more sophisticated form of the same dependency, financed nationally but still organized around the same productivity logic, is the open question the rest of this essay tries to answer. The deeper danger, in either case, is not that corporations will openly defy national interest. It is that national interest will quietly redefine itself to mean whatever keeps the infrastructure running — efficiency, uptime, compliance — rather than what it once meant: the cultivation of a population’s capacity to govern itself.
II. Developed and Emerging Nations: The New Architecture of Dependency
The asymmetry between developed and emerging economies in the AI era recapitulates a colonial trade pattern almost too precisely: raw material extracted cheaply from the periphery, refined at the center, sold back to the periphery at a markup it cannot itself produce. Where the old economy extracted cotton and returned cloth, the new one extracts behavioral data, language, and the unpaid or underpaid labor of populations who supply the training signal, and returns “intelligence” as a subscription.
The clearest documented instance of this pattern is Kenyan, not hypothetical. Beginning in late 2021, OpenAI contracted the outsourcing firm Sama to have workers in Nairobi read and label some of the most disturbing text on the internet — descriptions of child abuse, violence, and self-harm — so that the resulting safety classifier could filter such material out of ChatGPT before it ever reached a paying user elsewhere in the world. Workers were paid between roughly one and two dollars an hour for this; several later described lasting psychological harm, and some pursued legal action and a parliamentary petition in Kenya seeking protection and redress. The frontier model’s celebrated safety and the outsourced worker’s documented trauma are not two separate stories. They are the same transaction viewed from its two ends — value accruing at the center, cost absorbed at the periphery, with no mechanism by which the periphery shares in the valuation it made possible.
This is a difference between operational sight and architectural sight, and it mirrors a much older colonial educational policy, the kind that trained clerks fluent enough to run the empire’s paperwork but never scientists capable of redesigning its instruments. A population can be granted extensive permission to use a system and none at all to question its premises — and the two are easily mistaken for each other, especially by governments eager to claim digital modernity.
Here too, however, the picture is not static, and an honest political economy has to register the counter-movements alongside the extraction. Brazil’s central bank built PIX, an instant payments system now used by more than three-quarters of the adult population, almost entirely outside the rails of private global payment networks; its National AI Plan has committed several billion dollars to a domestic LLM ecosystem, including the Portuguese-language Sabiá model and a sovereign cloud initiative, while candidly acknowledging in its own planning documents that dependence on imported chips remains the binding constraint on genuine sovereignty. Indonesia is digitizing its social assistance rolls through a national digital identity system integrated with health and welfare data. India’s sovereign models, mentioned above, are a third instance of the same instinct. None of these efforts has yet closed the architectural gap; all of them confirm that emerging nations experience the asymmetry not as an abstraction but as a daily administrative fact, and are responding to it as such. Scale — billions of users, hundreds of languages, the demographic weight that India, Indonesia, Brazil, Nigeria, and Kenya together represent — is the asset most often invoked as the periphery’s counterweight, and it is real. But scale converts into leverage only through deliberate institution-building of exactly this kind; left to market dynamics alone, it converts instead into the largest possible captive market and the cheapest possible labeling workforce.
III. The Ultimate Consequence of Neoliberal Policy: From Institutional Capture to Ontological Capture
It is tempting, at this point, to reach for “neoliberalism” as an atmosphere — a generalized complaint about capitalism’s mood — rather than as a specific institutional history. That temptation should be resisted, because the more precise claim is also the more damning one: the particular shape AI has taken — privately owned, frontier-concentrated, optimized for engagement and productivity rather than designed as public infrastructure — is not an accident that befell a neutral technology. It is the direct institutional residue of four decades of deliberate restructuring across exactly the sectors that determine where transformative technology gets built and what it is built to do.
Begin with research itself. In the United States, the Bayh-Dole Act of 1980 allowed universities to patent and license inventions developed with public funding, converting a research commons into a portfolio of private intellectual property almost overnight. Over the following decades, the locus of frontier computing research drifted from publicly funded, mission-driven programs — the lineage that, through DARPA funding, produced the internet’s foundational protocols — toward venture-capital-financed startups and the research divisions of a handful of trillion-dollar firms. By the time transformer-based language models became technically feasible, the institutional capacity to build them at scale existed almost nowhere outside a small number of private companies, not because public science had become incapable of the insight, but because forty years of policy had systematically transferred the compute, the talent pipelines, and the commercialization rights away from public stewardship.
The same restructuring reached emerging nations through a different door: structural adjustment. Loan conditions imposed by the IMF and World Bank through the 1980s and 1990s required many Global South governments, India and Brazil among them, to cut public spending on education and health, privatize state enterprises, and reorganize remaining public services around cost-recovery and outcome metrics. This is why AI diagnostics and ed-tech tutors slot into many emerging-economy health and education systems with so little friction: the public alternative — a salaried doctor with unmetered time, a teacher accountable to a community rather than to a dashboard — had already been thinned out before any algorithm arrived. The same logic reaches the military domain by yet another route: the privatization of logistics, intelligence analysis, and now frontier-model defense contracts to firms like Anduril and Palantir, with OpenAI itself entering defense partnerships in the past two years, repeats at the level of national security the same transfer of capability from public, accountable institutions to private, commercially incentivized ones.
What unites research funding, education, health, and military procurement is not coincidence but a single institutional grammar: the conviction, installed across forty years of policy, that public stewardship is inefficient and private incentive is generative. AI is the sector in which this grammar reached its purest expression, because AI’s entire value proposition — efficiency, optimization, the elimination of friction — is the technological embodiment of the same logic that hollowed out the public sector. This is why the accusation belongs specifically to neoliberalism rather than to capitalism, technology, or modernity in general: it names an identifiable institutional project, with a beginning, a set of authors, and a list of sectors it deliberately reorganized, and AI is not external to that project but its most advanced product.
The same logic, applied to the self rather than to the institution, produces thinklessness. Neoliberalism’s defining achievement was never simply privatization; it was the conversion of the self into a unit of measurable productivity — what Foucault named homo economicus and what Byung-Chul Han has diagnosed as the achievement-subject, a self that exploits itself with the conviction that it is exercising freedom. AI does not disrupt this logic. It completes it, by offering the technology that logic had been waiting for: a means of outsourcing not just labor, but judgment itself, while preserving the appearance of individual agency. Once judgment is rented rather than exercised, citizenship contracts to what I have elsewhere called biological citizenship — the self reduced to its status as an eligible or ineligible recipient, a verified or flagged entry in someone else’s database. Arendt’s distinction between labor, work, and action locates exactly what is lost first: labor survives an AI economy easily, and may even be cushioned by it; action — the capacity to appear before others and initiate something genuinely new in public — atrophies, because action requires the very friction that AI’s commercial proposition exists to eliminate. The institutional capture described above and the ontological capture described here are not two different arguments. They are the same argument observed at two scales — the hollowing of public stewardship at the level of the nation, and the hollowing of judgment at the level of the person — and the second is simply what the first feels like from the inside.
A Bridge: Why Witness-Consciousness, Specifically
If the capture is ontological, the recourse cannot simply be a better set of metrics — a more humane productivity index, a gentler algorithm, a kinder dashboard. Any answer of that kind still operates inside the logic it claims to resist; it disputes what should be measured without disputing the authority of measurement itself, and a sufficiently sophisticated capture mechanism will eventually learn to optimize for “humane” too. What is needed is not a competing claim about what the self’s output should look like, but a demonstration that some part of the self is not output at all — not because the right instrument to measure it has not yet been built, but because witnessing and measuring belong to different orders of activity in the first place.
This is the precise sense in which the Upanishadic notion of witness-consciousness, the aham asmi that observes experience without being identical to any of its productive states, answers ontological capture rather than merely consoling against it. It is not a better-performing self; it is a structural remainder that no act of optimization can in principle reach, because it is defined as the observer of optimization rather than as one more thing optimization could improve. Where neoliberal AI capture works by making the self entirely co-extensive with its measurable output, witness-consciousness names exactly the part of selfhood that such an equation cannot complete — not a refuge from the productivity question, but a quiet refusal of its jurisdiction.
IV. Return to the Fundamentals of Humanity: A Possible Recourse
Why this resource, and not another
A reader rightly familiar with the wider Indian and South Asian intellectual inheritance will ask why witness-consciousness, drawn from the Upanishads, should be the philosophical anchor here rather than Buddhism, Gandhi, Tagore, Ambedkar, or Aurobindo — each of whom offers a substantial resource for resisting exactly this kind of capture. The honest answer is that they are not competitors so much as instruments tuned to different registers, and the choice made here is deliberate rather than exhaustive.
Buddhism’s doctrine of anatta, no-fixed-self, is in some ways a more radical answer than witness-consciousness: if there is no self to begin with, there is nothing for capture to seize. But this radicalism comes with a practical risk that contemporary history has already demonstrated. Buddhist-derived techniques — mindfulness, non-attachment, present-moment awareness — have been thoroughly absorbed into corporate wellness culture over the past fifteen years, repackaged as productivity tools that help employees tolerate intensified workloads rather than question them, a pattern critics have called “McMindfulness.” Any tradition organized around a technique one performs is vulnerable to exactly this re-absorption, because capture is very good at metabolizing practices. Witness-consciousness, as deployed here, is not a technique to be performed on a schedule; it is an ontological claim about what is already and always the case, which makes it considerably harder to package as a habit or sell as an app.
Gandhi’s swaraj and satyagraha operate at the register of political action and bodily discipline — they answer the question of what one does, and they are in fact already present in this essay’s emphasis on the panchayat and the public square as sites of resistance. But Gandhi’s framework presupposes a self capable of disciplined choice; it does not on its own explain what survives in a person whose capacity for judgment is being quietly captured before any choice is made. Gandhi and the argument made here are complementary, not rivals — one supplies the politics of recourse, the other its ontological ground.
Tagore’s critique of the mechanized, calculating “Nation,” and his defense of personality and creative freedom against organized civilization, sits closest in spirit to the argument here, operating at a civilizational and aesthetic register rather than a strictly philosophical one. He is less a substitute for this essay’s claim than its cultural correlate, and his suspicion of mechanism over personality could equally have furnished its epigraph.
Aurobindo is, if anything, the nearest ally rather than an alternative: his entire system of integral yoga is built directly on the Upanishadic witness, to which he adds an explicitly evolutionary dimension — the claim that recovering this layer of consciousness is not a retreat into a precolonial past but participation in an unfinished historical task. This essay’s recourse should be read in that spirit: not nostalgia, but an unfinished project resumed.
Ambedkar belongs in this list for a different and more disciplining reason than the others, and ignoring him would be a far more serious omission than choosing the Upanishads over any single rival tradition. Ambedkar’s method was critical retrieval, never wholesale inheritance: he broke from Hinduism precisely because he judged that its social order could not be separated from its texts, and he rejected Gandhi’s romanticization of the village republic as a den of localism and caste enclosure rather than a school of liberty. To draw on Upanishadic witness-consciousness honestly, in Ambedkar’s spirit rather than against it, requires explicitly setting aside the same textual tradition’s historical entanglement with Brahminical hierarchy and caste exclusivity. The philosophical proposition is portable; the social order in which it was first articulated is not, and does not need to be carried along with it. What this essay borrows from Ambedkar, more than a rival metaphysics, is this discipline of selective retrieval — and it is the same discipline the next section needs to apply to the institutions it goes on to recommend.
Building the recourse, institutionally
That discipline is necessary because the institutional recourse proposed for capture risks the very romanticization it should be guarding against. Arendt’s distinction between labor, work, and action suggests rebuilding small, real spaces where action survives — the gram sabha, the panchayat meeting, the railway platform conversation, sites where judgment is still exercised collectively and in person, where AI’s recommendations can be discussed, contested, and domesticated rather than passively absorbed. Sen’s capability approach offers the corresponding policy discipline: AI deployment should be evaluated not by niti, whether the rule was efficiently followed, but by nyay, whether the substantive freedom of the person in front of the system actually expanded. India’s Digital Public Infrastructure, and the parallel efforts visible in Brazil’s PIX and Indonesia’s digital identity rollout, gesture toward a third path between full corporate capture and full state control — though each, as their own planners admit, remains a vessel that inherits whichever logic is poured into it.
But the panchayat and the railway platform have already been invoked twice in this essay as sites of friction and deliberation, and honesty requires acknowledging what they have also been and often still are. Ambedkar’s own judgment on the village — not a republic of fraternity but, in his words, a structure of enclosed and reproduced caste hierarchy — should be taken at full weight here, not set aside for the sake of a tidy ending. Panchayats have been instruments of domination as often as forums of deliberation. Oral tradition transmits patriarchy and exclusion as fluently as it transmits wisdom. “Waiting” is not, for the woman kept outside a decision made by men in her own household, or for the petitioner standing for hours in front of a patwari’s desk, a contemplative discipline; it is the experience of subordination, a way the powerful make the powerless feel their dependency. “Ambiguity,” too, is not always depth — it is frequently the bureaucratic discretion that enables harassment and rent-seeking, the very thing that makes Bihar’s ration shops and block offices sites of extraction rather than only of philosophy. A reader who has watched any of this happen would be right to dismiss an ending that simply asked for a return to these forms as they have actually existed.
The recourse argued for here is therefore not a return to these institutions as substantively constituted, but the deliberate extraction of specific structural properties from them — friction, embodied co-presence, the necessity of persuading another person rather than satisfying an algorithm, the impossibility of instant resolution — while explicitly refusing the hierarchical content historically and presently attached to those forms. This is a harder task than nostalgia, because it is a design problem rather than a restoration project: building new sites that preserve the slowness and embodiment that made the old sites valuable, without inheriting the exclusions that made many of those same sites instruments of harm. Emerging nations’ deepest resource in the AI era may not be compute at all, but exactly this — populations who still remember, however unevenly and at however high a historical cost, what deliberation among present, accountable, embodied people looks like, and who therefore retain the standing to insist that whatever replaces it should be held to no lower a standard.
Closing
The geopolitical question with which this essay began — who owns the architecture of intelligence — turns out, on inspection, to be inseparable from an older question emerging nations have answered before under colonial rule, under structural adjustment, and now under algorithmic governance: whether independence secured at the level of the flag survives at the level of the self. The institutional history traced here suggests it does not survive automatically; forty years of neoliberal restructuring across research, education, health, and defense produced the very conditions under which AI now arrives pre-coded for extraction rather than for capability. The recourse available is neither technological catch-up alone, nor a sovereign LLM, nor a sentimental return to village India, but the harder work of building new institutions that hold both the structural conditions for human deliberation and a critical honesty about which inherited forms deserve to survive into them.
Comments
Post a Comment