Mitesh Khapra, an associate professor at IIT Madras, has been named on TIME’s list of the 100 most influential people in artificial intelligence. Unlike big tech leaders who dominate the list and work on global products, Khapra was recognised for his efforts to make AI accessible to India’s millions of non-English speakers.
With this inclusion, Khapra is in elite company – he rubs shoulders with the likes of OpenAI chief Sam Altman, Meta’s Mark Zuckerberg and xAI founder Elon Musk.
Closing the language gap in AI
Khapra has long argued that Indian language technology lags behind English because of a lack of quality datasets. To change this, he co-founded AI4Bharat, a research lab at IIT Madras dedicated to building open-source datasets, tools, and models for India’s many languages.
The team undertook one of the most ambitious projects of its kind – collecting thousands of hours of speech data from nearly 500 districts, spanning diverse educational and socioeconomic backgrounds. Their work now covers all 22 official Indian languages.
What is AI4Bharat?
AI4Bharat, a research lab at IIT Madras, is dedicated to advancing AI technology for Indian languages through open-source contributions. It was founded in 2019 by IIT researchers who wanted to create open-source AI models and datasets for Indian languages.
{{/usCountry}}AI4Bharat, a research lab at IIT Madras, is dedicated to advancing AI technology for Indian languages through open-source contributions. It was founded in 2019 by IIT researchers who wanted to create open-source AI models and datasets for Indian languages.
{{/usCountry}}Beyond technology, Khapra’s work is reshaping Indian academia. “Fifteen years back, an average PhD student in India working on language technology would end up working on English problems,” he was quoted as saying by TIME. “With these datasets available, I see a shift: now Indian students are working on Indian problems.”
From villages to the Supreme Court
AI4Bharat’s datasets today power almost every Indian startup working on voice technology. They also supply 80% of the data for the government’s Bhashini program, which aims to make digital services accessible in Indian languages.
The lab’s models are already in use across the country. They help the Supreme Court translate official documents and power a voice bot that farmers can call to report subsidy issues in their local language.
More about Mitesh Khapra
Khapra is an associate professor in the Department of Computer Science and Engineering at IIT Madras. Before academia, he was a researcher at IBM Research India, focusing on machine translation and deep learning.
He holds a PhD and MTech from IIT Bombay and has earned several prestigious honours, including the IBM PhD Fellowship, Microsoft Rising Star Award, and Google Faculty Research Award (2018).
His research has been published in top conferences such as ACL, NeurIPS, and AAAI, where he has also served as an area chair.