Five Things to Know About Mitesh Khapra's Work India has to build its own solutions instead of depending entirely on foreign technology, says Mitesh M. Khapra

In the Artificial Intelligence category, Time100 magazine highlighted global icons like Elon Musk, Sam Altman, Jensen Huang, and Mark Zuckerberg. What made the list notable for India, however, was the inclusion of Mitesh M. Khapra, an associate professor at IIT Madras, recognised for his pioneering contributions to developing AI for Indian languages.

Here are five key things to know about his work.

Building AI for India's many voices

Khapra heads the AI4Bharat Research Lab at IIT Madras which focuses on building datasets, tools, models and applications for Indian languages.

For him, the problem has always been clear. "The reason Indian language technology is behind English is because we do not have enough data for Indian languages," he shared with TIME.

To deal with this, Khapra and his team carried out one of the largest data collection efforts in the country. They went to nearly 500 districts, recording voices of people across different educational and social backgrounds. The aim was to capture the diversity of India's 22 official languages. These recordings later became open-source datasets that are now widely used by startups working on speech and language technology.

From research to real-world impact

The lab's work also connects to government efforts. AI4Bharat supplies nearly 80 per cent of the data for Bhashini, the government's program that allows citizens to access digital services in their own languages.

Also, its models have been deployed in the Supreme Court of India to translate documents, and in voice bots for farmers, who can now call in their native languages to resolve subsidy-related issues.

And Khapra isn't shy about the bigger picture, even if big tech companies use AI4Bharat's open-source datasets to improve their systems, it's still a win for India. "If big tech companies use our data to make their models better at Hindi or Marathi, it benefits the country at large."

A shift in academic culture

Beyond immediate applications, Khapra has been instrumental in reshaping India's AI research culture. "Fifteen years back, an average PhD student in India working on language technology would end up working on English problems," he says. "Now, with these datasets available, I see a shift, Indian students are working on Indian problems."

That cultural shift has positioned India not as a passive consumer of Western AI tools but as an active contributor to global research in multilingual AI.

Recognition and Responsibility

Khapra's work has brought him recognition at the highest levels from the Google Faculty Research Award (2018) to the Nasscom AI Game Changer Award (2021), and most recently, the Srimathi Marti Annapurna Gurunath Award for Excellence in Teaching (2022).

But for him, the bigger achievement lies in creating systems that can ensure India's digital sovereignty. His latest collaboration with Sarvam AI, a startup co-founded by his colleagues Pratyush Kumar from AI4Bharat focuses on building India's first homegrown foundation model for the government. Even if it falls short initially compared to Western systems, Khapra believes the effort is critical: "Unless we learn that skill, we will always be in a perpetually dependent position," he said.

His Vision

For Khapra, the journey has only just begun. His ambition is to bring Indian languages to parity with English in all AI-related tasks from sentiment analysis and speech recognition to more complex natural language generation.

"We want to build these solutions for this long tail of NLP tasks for as many Indian languages as possible," he shared with the media. It's a vision that is not just about technology but about inclusivity ensuring that every Indian, no matter which language they speak, has equal access to the digital future.