The government is prioritizing the development of home-grown voice-based AI models instead of relying solely on text-driven ones, which currently dominate the global AI landscape. This initiative aims to improve accessibility, allowing a vast segment of the population to engage with AI in their native languages through speech rather than text.
Unlike conventional AI chatbots such as OpenAI’s ChatGPT, Google’s Gemini, or Elon Musk’s Grok, which primarily operate on text, the government’s focus is on voice interactions. While global tech companies have integrated voice capabilities into their AI assistants, their models are predominantly optimized for English and a limited number of other languages.
The reason for this initiative is evident: most people find it easier to speak in their native language rather than type it. “The Internet will become more voice-enabled, and there will be many people who will prefer accessing services using voice commands,” Abhishek Singh, Additional Secretary at the Ministry of Electronics and IT (MeitY), stated in an interaction with FE.
Singh, spearheading the ₹10,372 crore IndiaAI mission, emphasized the market potential for locally developed voice-based AI models. “ If someone develops a voice-based large language model (LLM) in India, it will be a game-changer. Voice will become a significant enabler,” he said.
India has around 900 million active internet users, yet nearly 500 million people remain offline. Among them, rural communities, farmers, and those less familiar with technology may find spoken interaction with AI more accessible than written text.
For example, a farmer seeking guidance on crop diseases or weather updates could effortlessly ask an AI assistant in his native language rather than struggling with text-based interfaces. Likewise, individuals in need of health advice, information on government schemes, or educational resources could greatly benefit from AI-powered voice interactions.
The demand for India-specific voice AI models is substantial, as global AI companies have yet to fully address the country’s linguistic diversity. While some international AI systems offer multilingual support, their datasets for Indian languages remain limited. According to Singh, AI models trained on Indian datasets and tailored to the nation’s linguistic and cultural intricacies would outperform existing global alternatives in this space.
To fast-track the development of indigenous AI models, the government has received 67 proposals from startups, academic institutions, and private enterprises interested in building foundational AI systems. A technical committee is currently reviewing these submissions, with selected projects set to receive government funding.
“We have received around 22 proposals for large language models and 45 for smaller domain-specific models in sectors such as healthcare, education, and agriculture,” Singh said. The selection criteria will consider the teams' technical qualifications, the models' intended use, and the anticipated impact of their deployment.
A national dataset platform is also being developed to grant AI developers access to anonymized, non-personal data for training. This initiative seeks to establish a structured ecosystem where AI startups and researchers can effectively build, test, and enhance their voice-based models.
By focusing on voice-enabled models in various Indian languages, the government aims to make AI more inclusive and accessible to a broader population. This strategy not only helps bridge the digital divide but also strengthens the country's position as a leader in voice AI innovation.
Kommentare