AI will solve the issue of language inclusivity in India for DPI, says Nandan Nilekani

Nandan Nilekani, co-founder and non-executive chairman of Infosys

Nandan Nilekani, Non-Executive Chairman, Infosys, said that artificial intelligence (AI) can become the medium that will make digital public infrastructure-based (DPI) financial services accessible to billions of Indians.

India aims to take its DPI to over 50 countries in the next five years, and has already helped implement some parts of its DPI stack across 20 countries so far, Nilekani shared. He added that the estimated value of companies created on top of DPI will be over $100 billion in market cap.

Emphasising how AI will transform population scale services, Nilekani said that while AI is being used to reduce fraud by building highly personalised solutions, the use of AI to solve the language issue in India is also being considered. "We have 22 official languages and several hundred other spoken languages…So, some of the work we've done in AI for Bharat at IIT Madras, is to create open data for Indian languages, which can be used by everybody,” he added.

“I think AI has a huge role in communication,” he said at a fireside chat at Sahamati’s Samvaad 2024 held in Mumbai on November 19.

He then went on to point out that NPCI today is using these AI-driven language models to provide voice-activated commands in Hindi or English for payments. This has helped Indian users to speak in the language of their choice over the phone, or while using a WhatsApp bot or aggregating information to execute tasks over the phone.

Reiterating his thoughts on the usage of foreign-origin large language models (LLMs) by Indian startups and enterprises, Nilekani said that he continues to believe that LLMs are commodities, and the real value will unlock only with the layers and applications built on top of it or building the small language models using data specific to an enterprise or to a government.

“These four or five big guys are going to spend $200 billion to build the next LLM… Use cases are where the action is. I've always said that India will be the use case capital of AI in the world. So, you're going to see massive amount of AI in these applications, and some of them are going to be built by people,” he said.

Nilekani highlighted the 15-year-long journey of building India's technology stack of Aadhaar, UPI, DigiLocker, account aggregator infrastructure, and many more. “Over the last 15 years, there have been layers and layers of this (tech use cases). The important principle was to be population scale, enable small transactions, and frugal engineering. Making sure that we had public rates for private innovation. So right from day one, the philosophy has been, that we need these minimalistic rates. And then allow private innovation to flourish,” he said.