India can offer a blueprint for universal, trustworthy AI agents at scale

India should provide every citizen a free, privacy-first AI agent, built on digital public infrastructure, to ensure safe, inclusive access to information, services and decision-making tools

India is ready for the next leap in its digital journey. A decade ago, we created population-scale systems such as Aadhaar and the Unified Payments Interface (UPI) that made identity and payments accessible to everyone. These platforms worked because they were simple, secure, low-cost, and designed for universal use. As artificial intelligence (AI) enters daily life, we must bring the same clarity of purpose to a new public priority. Every Indian should have access to a safe and reliable AI agent that can help them manage information, navigate services, and strengthen their ability to make decisions.

 

For many people, an agent will become the main digital interface through which they receive information, learn new skills, and interact with both government and private services. A farmer may want timely insights on weather and soil. A student may need support for examinations or language practice. A patient may use an agent to understand prescriptions and organise health records. A small business may turn to an agent for logistics and compliance. These uses will only be realised if we build an inclusive architecture for all.

 

Four requirements should define an inclusive agent architecture. First, agents must comply with India’s privacy safeguards. Agents will handle sensitive data across health, finance, education, and daily behaviour. Users must know what information is collected and how it is used. Without transparency and trust, no system at the national scale will succeed. Second, agents must operate within India’s content guidelines. All societies set boundaries to reduce harmful content. India’s regulatory framework already defines these expectations.

 

Third, agents should be integrated with the Data Empowerment and Protection Architecture (Depa). Depa enables individuals to share their data securely through consent managers. When agents work with Depa, users retain control over when and how their data is shared. Finally, an agent should be available free of charge to every Indian with no advertising clutter. We have made UPI free; agents should be free as well. All systemically important digital intermediaries should therefore provide a free agent that meets safety and privacy norms. Companies can innovate above this layer with premium versions, but universal access must be guaranteed.

 

This approach is distinctively different from other jurisdictions. In the United States, AI agents are emerging as commercial products controlled by large platforms with high user surveillance. The European Union is focused on safety and rights but has not proposed an inclusive agent architecture. China is developing AI systems shaped by state priorities. None of these models is designed for a diverse democracy that must serve hundreds of millions of users with limited resources.

 

India can chart a different course because we already have the institutional foundations. Our digital public infrastructure is now one of the most complete in the world. Aadhaar provides secure identity. UPI delivers real-time payments to everyone. DigiLocker supports trusted digital documents. FASTag and CoWIN have shown how common standards can deliver public services efficiently. ONDC is building open networks for commerce and logistics. DigiYatra brings facial recognition-based air passenger flow and demonstrates how a privacy preserving, consent-based identity layer can improve service delivery. Depa adds a consent-based data architecture that even advanced economies are still debating. These systems function as a coherent stack connecting identity, payments, data, mobility, and service delivery.

 

An inclusive agent architecture would be genuinely transformative at a population scale. Across the world, access to high quality agent services is becoming linked to the ability to pay. If India guarantees universal access to well-regulated agents, we will demonstrate that AI can narrow inequalities. We would also offer a workable model for other low- and middle-income countries that cannot adopt American, European, or Chinese approaches. Many of these countries already look to India for digital public infrastructure. An inclusive agent system built around privacy, consent, openness, and public accountability would be of great value to them.

 

The case for an inclusive agent architecture is compelling. If AI becomes the primary way people interact with the digital world, then safety, affordability, and accountability must be guaranteed. Agents should support individuals across education, health care, agriculture, finance, and small business operations without creating new risks or dependencies. They should reflect our constitutional values of liberty, equality, justice, and fraternity, and work across India’s many languages and regions.

 

AI agents are likely to become routine digital offerings. The real question is whether their benefits will be available to everyone. The Agents for All framework offers a clear approach. If India moves with foresight, we can build an inclusive AI agent architecture that improves daily life for millions of people and sets a global benchmark for trustworthy AI.

The author is president, Everstone Group, and visiting professor in practice at the London School of Economics. He is a former Union minister and Lok Sabha MP. The views are personal