The Indian government has invested in the development of indigenous and sovereign artificial intelligence (AI) models to address key concerns around algorithmic bias and ensure that AI systems are grounded in Indian data and linguistic diversity, a senior official from the Ministry of Electronics and Information Technology (MeitY) said on Tuesday.
Speaking at the AI Impact Summit, MeitY Secretary S. Krishnan highlighted that building sovereign AI models was crucial for India’s technology strategy — not just from a capability standpoint but also to ensure that control over critical AI functions remains within the country.
He said sovereignty in AI means “nobody else has the kill switch, and it was always with us.”
Krishnan explained that while there was initial debate over whether to rely on open-source models developed abroad or build homegrown systems, the government opted for the latter to address specific challenges such as bias and linguistic representation.
“The first and foremost reason is to address the issue of bias — making sure that we have models primarily based on Indian data and Indian languages,” he said.
India’s linguistic diversity is enormous, with hundreds of spoken languages and dialects. According to Krishnan, it is important that AI tools are capable of understanding and interacting in local languages, a feature that global models often struggle with.
This includes voice-first and voice-text capabilities tailored to Indian users.
One of the flagship projects under this effort is BharatGen, a sovereign AI model launched in October 2024, which provides services such as Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) specifically for Indian languages.
The model aims to serve sectors including science, agriculture, manufacturing, and public services, empowering users to access AI systems in a familiar linguistic context.
Krishnan emphasised that the goal is not necessarily to compete with large global language models purely on scale, but to create useful, culturally relevant tools that can be widely adopted across India.
“It was important and critical to have a model of our own to truly represent India’s cultural and linguistic diversity,” he said.
The move reflects India’s broader push to build responsible and inclusive AI frameworks, ensuring that AI deployment is both locally relevant and ethical. With bias mitigation a priority, these homegrown models are expected to support a wide range of applications while reducing the risk of harmful or insensitive outputs that can emerge from models trained primarily on non-Indian datasets.
Bharat sarkar ne indigenous (desi) AI models develop karne me investment kiya hai taaki algorithmic bias jaise issues ko address kiya ja sake aur AI systems Indian data aur languages par based ho. Yeh baat Ministry of Electronics and Information Technology (MeitY) ke Secretary S. Krishnan ne kahi.
AI Impact Summit me bolte hue unhone kaha ki sovereign AI model banana strategic zarurat hai. Unka kehna tha ki AI sovereignty ka matlab hai ki critical technology par poora control India ke paas ho — kisi aur ke paas “kill switch” na ho.
Unhone yeh bhi bataya ki shuruaat me yeh debate thi ki India ko global open-source AI models use karne chahiye ya apne khud ke models develop karne chahiye. Aakhirkaar sarkar ne decide kiya ki India-specific challenges — jaise linguistic diversity aur bias — ko address karne ke liye desi models banana zaroori hai.
India me kai regional languages aur dialects boli jaati hain. Global AI models aksar Indian context aur local languages ko sahi tarah se represent nahi kar pate. Isi liye government ka focus aise AI systems par hai jo Indian data aur bhashao par trained ho, aur voice-based applications ko bhi support kare.
Is initiative ka ek major example BharatGen hai, jo October 2024 me launch hua tha. Yeh model Automatic Speech Recognition (ASR) aur Text-to-Speech (TTS) jaise features Indian languages me provide karta hai. Iska use agriculture, science, manufacturing aur public services jaise sectors me kiya ja sakta hai.
Officials ka maanna hai ki yeh step India ko responsible aur inclusive AI ecosystem build karne me madad karega. Long term me, indigenous AI models India ki digital sovereignty aur technological independence ko strong kar sakte hain.