New Delhi, February 17 Ethics will be a key factor in determining how effectively artificial intelligence (AI) is used in the healthcare sector, said Jitendra Singh, Minister of State (Independent Charge) for Science and Technology and Earth Sciences, on Tuesday.
Citing the example of the government's sovereign multilingual AI engine, BharatGen, he said that the current government is prepared to take on new initiatives, even at their early stages.
"Ethics will be a very important factor in determining how optimally AI can be used in the healthcare sector," the minister said.
He was speaking at a session at the AI Impact Summit.
Singh said that BharatGen is a government-owned sovereign model, but there are also other private sector players working on similar models, such as digital health and digital case histories.
"In the future, we will also need to integrate. Neither can we work in isolation, nor can they. So, we need to be prepared for that," he said.
Launched in October 2024, BharatGen is a government project to develop a sovereign AI model that will provide services such as Automatic Speech Recognition (ASR) models and Text-to-Speech (TTS) models for Indian languages.
At the session, BharatGen unveiled its new 17B (billion) parameter LLM (large language models).
Currently, BharatGen's AI models support several Indian languages, including Hindi, Assamese, Bengali, Gujarati, Kannada, Maithili, Malayalam, Marathi, Oriya, Punjabi, Sanskrit, Sindhi, Tamil, and Telugu.
BharatGen has released domain-specific fine-tuned models for Ayurveda (Ayur Param), Indian agriculture (Agri Param), and the Indian legal domain (Legal Param). In addition, all BharatGen models (text, speech, and vision) are useful for applications across healthcare, agriculture, education, and governance.
The 17-billion-parameter multilingual AI model, called Param2, will support 22 Indian languages. The model is developed under the BharatGen consortium, which operates out of the Technology Innovation Hub at IIT Bombay and is supported by the Department of Science and Technology.
Rishi Bal, CEO of BharatGen, said that Param2 is a 'mixture of experts' model built to work across 22 Indian languages.
Highlighting the technological depth of the project, the minister noted that BharatGen spans multiple AI modalities, including text-based large language models, speech technologies such as text-to-speech and automatic speech recognition, and document vision-language models.
BharatGen's foundational models are designed for inclusive, India-centric applications in governance, healthcare, education, agriculture, and legal systems, especially in linguistically diverse regions, he added.
Prime Minister Narendra Modi has launched BharatGen's latest models, including the Param2 text foundation model in 22 scheduled Indian languages, with 17 billion parameters, Shrutam speech-to-text models in 12 Indian languages, Sooktam text-to-speech models in 12 languages, and Patram models under the DocBodh framework for multilingual access to complex Indian documents.
Singh emphasised that India's linguistic diversity extends beyond the 22 scheduled languages, and stressed the need to continuously expand data sets to include widely spoken regional languages and dialects.
