India is building population-scale intelligence through voice AI: EkStep Foundation chair Nilekani

Bengaluru: India has taken Artificial Intelligence (AI) beyond the hype and hand-waving, to actually making things on the ground, said Nandan Nilekani, co-founder and chairperson of EkStep Foundation.

Speaking at an event here on Wednesday, organised by EkStep on ‘Voice AI — Making the Best Work for India’, Nilekani said the final frontier of access is voice. “If a person can talk to the computer and receive instructions or information or get an agent to work, that’s the final concept, and even more so in a country which has so many different languages and dialects,” the co-founder of Infosys said.

The event saw many startups showcasing their Indic-language voice AI deployments. Nilekani said there is a race in AI, a race to the bottom and a race to the top. Race to the bottom includes many AI-generated contents, including pornography, and others that affect people’s mental health.

“Race to the top is how do we use AI to improve people’s lives? We want to make that race to the top happen, because that will actually help the AI community live up to the expectations that are being set,” he said, stressing that one can excel in building intelligence at a population scale, especially in voice AI.

Speaking to Nilekani in a fireside chat, Vishal Dhupar, MD of South Asia, NVIDIA, said that with the work that is being done in the country, we are going to become the capital of intelligence.

With people coming here and taking advantage of the talent, Dhupar said he sees a 3X kind of an opportunity for India in the next 3-4 years.

Many startups are now focusing on voice AI. With many dialects and languages in India, voice AI provides multilingual support across vertical, be it health, railways or HR-related functions.

Nilekani said voice AI is not a competitive space and that we are not competing on it. “Everybody will use voice AI as a utility to improve things, and then we’ll share learning. Voice is a whole new frontier, simply because there’s so much variation, so many dialects. We speak English or English words liberally between languages that AI should deal with and be able to comprehend and answer,” he said, adding that AI is about enhancing human beings.