(Bloomberg) — Follow Bloomberg India on WhatsApp for exclusive content and analysis on billionaires, companies and market trends. Sign up here.
Most Read Articles on Bloomberg
A consortium backed by Mukesh Ambani’s Reliance Industries Ltd and India’s top engineering schools aims to launch its first ChatGPT-style service next month, a move that could boost the country’s ambitions to become a player in the artificial intelligence space. It’s a big step.
The BharatGPT group, which comprises India’s most valuable companies and eight partner universities, unveiled an overview of its large-scale language model during a technology conference in Mumbai on Tuesday. Videos played in front of attendees showed a motorcycle mechanic in southern India asking the AI bot questions in his native Tamil, a banker talking to the tool in Hindi, and a developer in Hyderabad asking the AI bot questions in his native Tamil. The video showed him using it to write computer code.
If successful, the model (named Hanuman after the half-monkey Hindu god) would represent a step forward for India in the race to develop potentially transformative AI technology. BharatGPT envisions a model that will work through 11 local languages in four key areas: healthcare, governance, financial services and education. The company developed the model in collaboration with Indian Institutes of Technology in Bombay and elsewhere, with support from wireless carrier Reliance Jio Infocomm Ltd. and the Indian government.
Startups like Sarvam and Krutrim, backed by prominent VC investors such as Lightspeed Venture Partners and billionaire Vinod Khosla’s fund, are also building open-source AI models customized for India. Masu. Silicon Valley companies like OpenAI are building LLMs at increasingly large scale, but their efforts are hampered by computational constraints and simpler models available to small businesses and government sectors. A plan is needed.
“This is a different genre of LLM,” says Ganesh Ramakrishnan, dean of the Department of Computer Science and Engineering at IIT Bombay.
Read more: Ambani aims to complete Reliance’s AI transformation in 2024
“Hanuman will also provide voice-to-text conversion functionality, making it significantly more user-friendly,” Hanuman said in an interview on the sidelines of the annual IT industry conference Nascom. In a country of 1.4 billion people, millions of people are illiterate.
Reliance Jio will build customized models for specific applications, he said. The telecom-to-retail conglomerate is already working on his Jio Brain, a platform that uses his AI across its network of around 450 million subscribers.
LLM is a system that learns from vast amounts of data and generates natural responses. Such models utilize generative AI, a new type of artificial intelligence popularized by the success of OpenAI’s ChatGPT.
BharatGPT’s efforts in this area are somewhat unique. This is the first public-private partnership of its kind in the country, involving leading companies from different fields.
“It’s like the joint family in India,” Ramakrishnan said, referring to the intergenerational family structure that is still common in the country. “We are interdependent and we do better together.”
Read more: Ola founder Krutrim becomes India’s first $1 billion AI startup
Most Read Articles on Bloomberg Businessweek
©2024 Bloomberg LP
