ADVERTISEMENT

As AI technology evolves faster than it is being adopted, Nandan Nilekani, Co-founder and Chairman of Infosys , said at a fireside chat during the India AI Impact Summit that India’s experience in building large-scale public digital infrastructure would help it identify the right path for AI diffusion at scale.
“We learned that diffusion is a technique. It’s both art and science. It involves institutions, policymaking, negotiations, dealing with incumbents and newcomers, and strategies for execution. If all the investments in AI are to deliver value to society — not just to individuals — we will have to focus on diffusion pathways to take this to everyone,” he said, adding that India would lead on that front. “That’s why I’ve always said India should focus on becoming the use-case capital of the world.”
Citing the example of how India’s digital public infrastructure model has expanded globally to nearly 15 countries, Nilekani said India possesses the essential components — a strong political commitment, dedicated technologists, and extensive experience in large-scale technology deployment — to take the lead in the successful implementation of AI. “India is a country that is very positive about technology in general and AI in particular. We need to take advantage of that optimism and not let people down by delivering truly transformative AI applications, which you will see in the next two to three years,” he said.
ALSO READ | FedEx plans AI models to predict supply chain vulnerabilities due to trade, other disruptions
Emphasising the need for AI diffusion to be inclusive, Nilekani said technology must be able to function seamlessly across languages. “We want people to be able to speak to the computer in their own language, in their dialect — mixing English, Hindi, Tamil or any other language. That needs to be done, and it should be done. I think that’s a big step,” he said.
Highlighting the role of AI agents, he added, “If you can make agents work for people, it leads to greater inclusion because complex tasks can be handled while the sophistication remains hidden behind the agent.”