The Need for Building AI for Bharat

As the world grapples with the impact China’s DeepSeek has created, Indian companies have included the model in their offerings, balancing the race and also driving innovation in the country. But the question remains: is this enough when it comes to building AI solutions that cater to the needs of the country’s larger population?

Speaking with AIM at the podcast What’s the Point, Shekar Sivasubramanian, CEO of Wadhwani AI, explained the exact definition and the differentiation between AI for India and AI for Bharat.

There have always been two distinct Indias within India. In the current context, we can interpret them as the urban, tech-savvy ‘India’ and the rural ‘Bharat’ with its first-time smartphone users. This divide presents unique challenges in bridging the gap between technology and its diverse users.

“The tech that is built around the world is mostly focused on India,” Sivasubramanian said. “It already assumes a certain amount of familiarity, knowledge, education, and understanding of technologies that have been used thus far.”

Delivering to this community is easy. However, the challenges become more complex when it comes to delivering AI solutions or any technology to someone using a smartphone for the first time. For this population, the experience is entirely new, and as a result, there is a lack of trust, knowledge, and contextualisation of what they get in their hands – in this case, an AI chatbot.

“To deliver to a farmer, you need to know what their free time in the afternoon is and where they will use technology. You never think of that when you’re delivering in India,” Sivasubramanian said. According to him, this is also coupled with the fact that a section of the Indian population consists of people who speak thousands of dialects and languages and are not accustomed to English.

The Need Becomes Clear

Nikhil Malhotra, chief innovation officer at Tech Mahindra, told AIM that his definition of ‘AI for Bharat’ aligns with Sivasubramanian’s. While building Project Indus for Tech Mahindra, Malhotra and his team also visited different cities and towns in India to collect data for training models in Indic languages. 

“This is important because for anyone in India, their first thought process triggers…in their own language – that’s AI for Bharat for me,” Malhotra explained, emphasising that this goes beyond people who think natively in English. He added that this includes not just the rural population but also many homemakers in the country who are first-time smartphone users.

For Ankush Sabharwal, founder and CEO of CoRover and BharatGPT, AI for Bharat means not just introducing AI to Bharat but also using AI to educate the Indian population about government policies, changes, and schemes.

“We are what we are,” Sabharwal said, emphasising that while educating the population in English and other skills is valuable, language should never be a barrier to access and opportunity. “Make everything natural and accessible [through AI].”

This is where voice-enabled AI plays a crucial role. Malhotra explained that since most people naturally prefer speaking over typing or writing, building speech models tailored for the Bharatiya population becomes important.

Similarly, Malhotra said that India’s 5,000-year-old civilisation holds a wealth of valuable information that can be used to build better AI solutions for the population. He cited an example of how his team built an algorithm for Panchang, which not only works in India for accurately predicting the weather but also abroad, particularly in Sydney. 

This is also why teaching AI in schools becomes extremely important. Realising this, the CBSE has also introduced AI as an optional subject in schools. Sivasubramanian noted that the curriculum is so advanced that even experts in the AI field right now will find it challenging to complete.

Building Trust is the Most Important Thing

However, building public trust in AI is easier said than done. Sivasubramanian explained this with an example: if farmers are asked to click pictures of pesticides, they often capture not just the pesticide but also pictures of the surrounding area, including the sky, trees, and even their friends.

“But you cannot restrict or constrain that. That is how they will learn and investigate the technology. We should help them instead of criticising them,” Sivasubramanian said, adding that the most important thing that researchers should focus on is building trust with people. 

There is no better measure of the success of a technology than its acceptance by the people it aims to serve. “Keep everything else aside,” Sivasubramanian said. He added that the most important way to do this is to talk to people with empathy and vulnerability. 

“We don’t talk down to them just because we are technologists. We understand that their dignity and work are as important as ours. We sit down one-on-one with them as equals and chat and all the truths and problems come out.”

Sivasubramanian cited another example from when a researcher from Wadhwani AI was trying to explain an AI chatbot to a farmer who was not very interested in learning about it. After listening for a few minutes to the researcher, the farmer just said, “I do not understand whatever you are saying, but I understand that you care about me.”

The researcher thought he failed to explain the solution to the farmer. However, Sivasubramanian saw it differently. He said this was actually the researcher’s greatest success because he had built a bridge of understanding.

“The opportunity to serve somebody is the only gift that is long-lasting. Not fame, not money, not designation, not power,” Sivasubramanian said. He believes the goal should not be earning billions of dollars but, instead, helping somebody and making their lives better.