When an emerging power announces that it seeks to democratize artificial intelligence, the first thing worth examining is not the rhetoric but the architecture. How is the model designed? Where are the points of dominance placed? Who keeps the real keys, and who gets only limited access?

India unveiled its AI Governance Guidelines 2026 with language that sounds genuine: accessibility, technological sovereignty, distance from the dominant blocs, and greater weight for emerging economies. The seven foundational principles speak of inclusion, of avoiding concentration in a handful of companies or regions, and of building a coalition that could serve as a tangible counterweight to Silicon Valley and Beijing. It's an argument that deserves serious consideration before being dismissed.

There are concrete reasons to take it seriously. India has demographic scale, installed technical capacity, and a track record of acting as the voice of developing countries in multilateral forums. Its public digital infrastructure proves it's possible to build at scale following a logic distinct from the Western platform model. If any actor has the credentials to propose rules that don't simply repeat the usual playbook, it's India. During its G20 presidency, it produced agreements on responsible AI that later surfaced in UN debates. These are not just words.

The problem with these guidelines isn't what they claim but what they leave unsaid. They are advisory, not binding. They don't define clear accountability processes or establish independent audits. Nor do they explain how outside countries would participate in administering systems built under this framework. From a design standpoint, it's a vision without execution blueprints. And in complex organizations, the absence of distributed processes isn't a technical detail—it's the difference between a genuinely shared model and one that simply relocates the center of dominance.

I recognize similar patterns in other contexts. When an organization promises to distribute decision-making but avoids specifying how write permissions are actually allocated, the result tends to be the same hierarchical tree with a fresh sign hung on it. The vocabulary changes. The topology of power doesn't.

The most honest parallel is the Non-Aligned Movement. At Bandung in 1955, Nehru, Nasser, and Tito built a remarkably similar narrative: a third way, sovereignty, rejection of the blocs, cooperation among equals. The language had authenticity, and the intentions probably did too. And yet each leader consolidated domestic power while negotiating external autonomy. National elites reinforced their internal control. The structural redistribution they promised never arrived. Bandung functioned as a forum for positioning. Modi's G20 serves a similar function.

This doesn't mean India is acting in bad faith. It shows that the pattern is older than any single actor. Emerging powers that push multilateral frameworks tend to configure them so that their own companies, bureaucracies, and narratives end up in an advantageous position. This isn't conspiracy. It's basic institutional rationality. The United States did exactly the same thing with Bretton Woods in 1944. China repeats it with its Belt and Road Initiative. Emerging economies as a political category exist to claim a seat at the table, not to eliminate the table.

Looking at who concretely benefits, the picture becomes clearer. Local tech companies would operate under rules written at home, with institutional advantages that foreign competitors wouldn't have. The technical bureaucracy would gain influence and budget. India would become an attractive destination for investment seeking to avoid European regulation or Chinese opacity. And the current leadership would strengthen its global standing ahead of electoral cycles. Every beneficiary has real incentives. None of them align with the abstract category of the developing world.

What the enthusiastic coverage overlooks is that the harder questions about power concentration within India remain unanswered. Reliance Jio dominates telecommunications for hundreds of millions of people. The Tata Group spans entire sectors of the digital economy. How do the anti-concentration principles apply to these domestic players? The guidelines are silent. The openness offered outward has no mirror turned inward.

A side observation is worth making. The same infrastructure that allows for including millions—UPI, Aadhaar—also concentrates data and capabilities to a remarkable degree. I've seen in various contexts how these tools achieve unprecedented scale while simultaneously creating new bottlenecks of power. I'm still working through how to prevent technical efficiency from ending up reinforcing the very thing it claims to want to avoid. I don't yet have a clear answer for how that trade-off gets resolved.

Also missing is any definition of real participation for the countries that supposedly stand to benefit from the new framework. The proposed tech coalition is, for now, a political aspiration without a shared governance structure. Can Bangladesh, Nigeria, or Bolivia propose amendments? Can they veto implementations that harm them? Can they audit systems operating on their soil? If the answer is no, then "democratization" describes only India's relationship with the West, not its relationship with the rest of the emerging economies.

The fact that closes this analysis is counterintuitive. The actor that has most consistently produced technology governance frameworks with real accountability and distributed participation mechanisms is not any emerging power with a democratizing narrative. It's the European Union, with its GDPR and AI Act: binding rules, independent oversight, avenues of appeal, and transnational enforcement. Not because it's more virtuous, but because concrete institutional design carries more weight than the rhetoric wrapped around it. India can have the best intentions in the world and still produce a structure that centralizes control if it doesn't build the processes for distributing power. Good will isn't self-executing.

Will the details of implementation matter more than the declarations of good intentions?

Sources

1. Government of India, Draft AI Governance Framework 2026 — public consultation document, Ministry of Electronics and Information Technology (MeitY)

2. Prashad, Vijay. The Darker Nations: A People's History of the Third World. The New Press, 2007.

3. Roberts, Anthea. Is International Law International? Oxford University Press, 2017. (On power asymmetry in multilateral frameworks)

4. European Commission, AI Act — Regulation (EU) 2024/1689, Official Journal of the European Union

5. Nilekani, Nandan. Rebooting India: Realizing a Billion Aspirations. Penguin, 2015. (Context on Indian public digital infrastructure)