An intelligent way to ensure AI sovereignty

India's AI opportunity is not to win the race to the most advanced model.

Key points

• Urvashi Aneja Founder and Director, Digital Futures Lab and Co-Convenor, Global South Network for Trustworthy AI.
• The light-touch approach of the November 2025 AI Governance Guidelines makes this more urgent: rules that rely on principles and self-regulation only work alongside continuous evidence generation.
• India's ecosystem of context- and domain-specific models is genuinely vibrant, but almost all of it sits on top of a handful of American labs.
• Serious work is now under way in India on model performance, including multilingual benchmarking.
• Throughout much of the West, enthusiasm for AI has given way to suspicion.

India's AI opportunity is not to win the race to the most advanced model. It is to show the world how AI can be created useful to real residents, solving real challenges. That is what India's own strategy notes, and what framed the 2026 India AI Impact Summit. India has the ingredients: linguistic and cultural diversity, digital public infrastructure at scale, a young technical workforce, and startups working on challenges that matter. Two things stand in the way.

Notably, the first is that we are measuring impact the wrong way. In the current policy discourse, impact has come to mean adoption and diffusion-how plenty of sectors, how numerous users, how plenty of languages. These are useful numbers, but they say nothing regarding whether the systems being adopted actually deliver, and nothing concerning the new vulnerabilities they create. A system that performs adequately where institutional oversight is solid can cause serious harm where that oversight is thin. For context, the same conditions that let India leapfrog additionally leave it exposed. Impact has two halves; we are counting one.

In practice, the second is dependency. India's ecosystem of context- and domain-specific models is genuinely vibrant, but almost all of it sits on top of a handful of American labs. Dependency in itself is not the difficulty; it exists in numerous sectors. It becomes a difficulty when you cannot negotiate its terms. As seen in agreements being signed between leading AI firms and global philanthropies, AI credits have meanwhile become a form of development assistance.

Neither difficulty has a single fix. Both will need competition policy, public compute, better redress mechanisms and much else. But one piece is common to both, and it is the piece India has invested in least: independent evaluation.

To define impact fully, India must be able to see how the systems it adopts actually perform throughout its languages and settings, and to detect the harms adoption produces. To manage dependency, it must be able to set conditions on what it accepts, and to act when those conditions are not met. Both turn on the same capacity.

This makes evaluation strategic rather than merely technical. It is the layer at which India can assert sovereign authority even where the underlying models are not sovereign. Sovereignty in AI need not mean building the models. It can mean choosing between them on terms India has set.

Procurement is where this becomes real. The Indian state is unusual in the scale of its role as an AI adopter: among the largest purchasers, a builder of public compute and datasets, and a funder of indigenous foundation models. A procurement regime that requires independent evaluation-demonstrable performance in Indian languages, evidence of safety in the deployment context, post-deployment monitoring for high-stakes uses-turns the State's buying power into a tool for raising market standards. The light-touch approach of the November 2025 AI Governance Guidelines makes this more urgent: rules that rely on principles and self-regulation only work alongside continuous evidence generation.

None of this is straightforward, and the field is far from settled. A systematic review of 445 large-language model benchmarks discovered only 16 percent employed rigorous methods to compare model performance, and roughly half asserted to measure abstract qualities like "reasoning" or "harmlessness" without defining them. The challenges compound throughout languages: test datasets do not exist for plenty of low-resource languages, and translating English ones fails to capture how residents speak or prompt.

Since the most consequential effects of AI are not visible at the model level, even good benchmarks are not enough. They emerge as individuals apply systems over time. A frontline health worker who routinely defers to an AI recommendation may lose judgement she once exercised. A welfare officer who overrides an algorithmic flag may face pressure to do so less often. Beyond these interaction harms sit systemic ones-effects on labour markets, on human agency, on cognitive ability.

This is where the gap is starkest. Serious work is now under way in India on model performance, including multilingual benchmarking. On interaction and systemic harms, there is almost nothing-a global survey of generative AI evaluations discovered fewer than 6 percent accounted for human-AI interaction. These harms need observational, longitudinal, often qualitative methods, and domain knowledge more than technical knowledge. They are slower, costlier and harder to reduce to clean numbers. It is why they are barely funded.

Public trust is additionally a factor. Throughout much of the West, enthusiasm for AI has given way to suspicion. India is in a different position: public sentiment here remains optimistic regarding technology. That optimism is an asset, but it is not permanent. It will hold only if these systems prove safe as well as useful, and if failures are caught by someone other than the firms that built them. Evaluation is how that gets demonstrated.

India should additionally choose its niche. Frontier capability testing is already being done in the UK and the US. It have the compute and secure infrastructure for it; there, India needs partnerships for joint testing and timely sharing of results. Its comparative advantage lies in the application layer, multilingual evaluation and post-deployment monitoring-work every country in the Global South will need. It makes it exportable.

Three steps would obtain us kicked off. As Singapore does, where third-party evaluators test commercial AI applications and the state turns findings into public guidance, but built around domain communities of practice, with doctors, teachers, farmers and legal aid workers designing the rubrics, fund application-layer sandboxes. Focus the AI Centres of Excellence in health, agriculture and education as certification and assurance bodies rather than pushing them only towards commercial products; assurance is the clearer public-interest mandate.

Third, and most neglected: grow an ecosystem of independent third-party evaluators. None of this works without individuals to do the work, and India today has highly little capacity-some at technical institutions, some in civil society, nothing near what the task requires. The field has to be actively cultivated: training, shared methods and tooling, and institutional homes where an evaluation career is feasible. It should be a named budget line for governments and development finance agencies funding AI adoption, not an afterthought.

Civil society organisations belong at the centre of that ecosystem, not at its edges. They see systems in employ, in context, over time-a vantage no laboratory has, and exactly what interaction and systemic harms require. Their absence from formal evaluation notes more concerning how we have defined technical expertise than regarding what the work demands.

Finally, evaluation is labour. As content moderation does, building datasets and running safety tests exposes individuals to disturbing content. Fair pay, psychological backing and transparent contracting belong in any publicly funded evaluation programme.

India wants to be the 'AI employ capital' of the world. It cannot be that without additionally becoming its evaluation capital.

Taken together, the developments around an intelligent way to ensure AI sovereignty point to a situation that is still moving, and the coming days should bring more clarity.