India uniquely poised to combine sprawling DPI with power of AI: Nandan Nilekani

India is uniquely placed to combine its sprawling digital public infrastructure (DPI) with the power of artificial intelligence to unleash transformative use cases, Infosys chairman and Aadhaar architect Nandan Nilekani said on Friday.

Speaking at Carnegie India's Global Technology Summit 'Sambhavna', Nilekani cautioned that while artificial intelligence (AI) hype cycles have reached unprecedented levels, real challenges exist in building and operationalising it at scale.

The running theme in India is how to use AI to make lives better at a low cost and at a population scale. "AI makes DPI better...The digital infrastructure we have is getting loaded with AI, and making DPI better...India will be uniquely placed because of its history, and it will combine DPI and AI to create a whole new way of doing things," Nilekani said.

The AI playbook in India is all about focusing on individual narrow use cases, and ensuring it is safe, secure, unbiased, and responsible.

AI is not easy, certainly not an overnight "magic", he said with a note of caution.

"There are the usual issues. How do you make the user experience seamless? What is the infrastructure we need? How do you govern these things? How do you manage the complexity? These are things we have dealt with in every field, and we are dealing with them at AI at scale," he said.

While many of these challenges in AI are typical of new innovations in every field, the difference between previous tech advances and this one is that for the first time, AI requires trust to be placed in non-human intelligence for decision-making.

"...we didn't do that earlier because previous technology was deterministic, predictable. Now, we are essentially expecting the machine to make decisions and there's a huge leap of confidence, a huge leap of faith in the ability of technology to take us forward," he said.

While people tend to be far more forgiving of human error, the same is not the case when one is dealing with a machine error. A case in point, he said, is autonomous cars, where any fatality on the road would force the provider to go back to the drawing board for two years.

Adopting AI at scale is hard work and will continue to be so, he said, adding that an even bigger challenge is the adoption of AI at scale in enterprises and public sector space.

"There's a lot to be done if you really want to make all this AI stuff work. And all transitions are painful...This has happened before, in every industry...The difference this time...is the hype is at a different level. And people think it is some 'pixie dust'," he said.

Nilekani emphasised that AI adoption will not be easy.

"This is something which is much more complicated than we think," he said. The industry veteran further observed that adoption cycles are reducing globally in general.

"Each cycle of technology takes less and less years. And in general, it has taken so much more time to get adopted in India. Interestingly, this time around, while we expect AI adoption also to take 10-15 years, our belief is that in India it can happen much faster," he said.

The tech transformation in India would enable the country to leapfrog and narrow the gap between global developments and AI build-up here.

Pointing to the rise of India's digital infrastructure, he said Aadhaar and UPI led to an explosion in the payments and transaction ecosystem.

As a result, the balance moved from just the global tech companies to homegrown ones backed by venture capital, Nilekani said as he cited examples of companies like Meesho, PhonePe, Physics wallah, Zepto, Rapido and others who redefined the spaces with their innovative offerings.

"India's DPI will be the basis for using AI. So, when we look at implementing AI at scale in India, it is built on an existing foundation of digital transformation that has happened for a decade," he said.

Going forward, the local languages, and user interface moving to voice and video (from keyboard and touch) would drive AI adoption and accessibility to a billion people.

"We believe this will lead to India becoming AI use capital of the world...DeepSeek or not, India will be the place where stuff (AI) gets used at the scale of one billion people like we have shown before with Aadhaar and UPI," Nilekani pointed out.

Today, innovative companies are working on open-source AI models for Indian languages, reducing the cost of operation. Exorbitant prices won't work here, he said advocating low-cost, population-scale AI.

Much of AI is about data, and its organic collection, not about building models by scraping the internet and someone else's IP, he asserted.

The focus in India is to use AI to make lives better, not to make "things convenient".

"It is not how do we use AI to make things so convenient that you lose your skills. It is not about dumbing down people, it is about using AI to improve the capacity and potential of human beings," he said.