BharatGen to complete AI models based on 22 languages this month: Jitendra Singh
The government’s sovereign multilingual AI engine BharatGen is expected to complete text-based service by the end of this month, Union Minister Jitendra Singh said in the Rajya Sabha on Thursday.
Replying to a supplementary question by BJP member Bhubaneswar Kalita, the Minister of State (Independent Charge) for Science and Technology said the work for developing BharatGen AI is a dynamic process and may add more languages as well as dialects.
“We have already completed 15 languages. We will be completing 22 (official languages) in this month itself. So all the 22 will have text completed by this month and 15 will also have the speech and the vision module,” Singh said.
The minister said the BharatGen initiative is being executed through a network of 25 Technology Innovation Hubs (TIHs), four of which have been upgraded to Technology Translational Research Parks (TTRPs).
In response to a question by BJD member Sasmit Patra on data sharing of BharatGen with Indian innovators and pricing, Singh said the data will be shared with certain terms and conditions.
He said discussions are on for offering BharatGen service at discounted price.
AAP member Raghav Chadha asked about predictable access to GPU (graphics processing unit) to meet computational requirements for the development of artificial intelligence (AI) technology.
In his reply, Singh said that GPU is at the core of the entire exercise around AI development.
The minister said larger AI models may require thousands of GPUs and access to computers is being granted through IndiaAI mission.
“We are in an evolving mode. The mission has just started but we are very aware and cautious of challenges. As far as funding is concerned, I think one good initiative about this government is that we have opened up to the private sector now. We are also now encouraging more investments from non-government entities,” Singh said.
Launched in October 2024, BharatGen is the first government supported national initiative to develop a range of sovereign foundational AI models tailored to Indian languages and societal context.
It will provide multiple services, including text (via Large Language Models), speech (Text-to-Speech and Automatic Speech Recognition), and vision-language systems.
Currently, BharatGen’s AI models support 15 Indian languages, including Hindi, Assamese, Bengali, Gujarati, Kannada, Maithili, Malayalam, Marathi, Nepali, Odia, Punjabi, Sanskrit, Sindhi, Tamil and Telugu.
