BharatGen: India’s First Government-Funded Multimodal AI Initiative

On September 30, 2024, BharatGen was inaugurated in New Delhi, marking a significant step in India’s commitment to generative AI. The initiative aims to enhance public service delivery and citizen engagement by developing foundational models in language, speech, and computer vision. Dr. Jitendra Singh, Union Minister of State for various portfolios, emphasized BharatGen’s role in positioning India as a global leader in generative AI, similar to the transformative impact of UPI.

Key Objectives

BharatGen, led by IIT Bombay under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS), focuses on creating generative AI systems that produce high-quality text and multimodal content in various Indian languages. The project is implemented by the TIH Foundation for IoT and IOE at IIT Bombay, with contributions from several premier institutions, including IIIT Hyderabad and IIM Indore.

Core Features

The initiative is distinguished by four key features:

Multilingual and Multimodal Models: BharatGen will create foundational models that cater to diverse Indian languages and formats.

Bhartiya Dataset: The project will prioritize the development and training of AI systems using India-centric datasets.

Open-Source Platform: BharatGen will promote an open-source ecosystem for generative AI research in India.

Focus on Data Sovereignty: The initiative emphasizes collecting and curating data that accurately represents India’s linguistic and cultural diversity.

Social Impact

BharatGen is designed to ensure that generative AI serves as a public good, promoting social equity and cultural preservation. The project aims to make AI accessible across various segments of society, addressing national priorities such as inclusive technology development.

Atmanirbhar Bharat Vision

Aligned with the Atmanirbhar Bharat initiative, BharatGen seeks to develop foundational AI models tailored specifically for India, reducing reliance on foreign technologies. By fostering a robust domestic AI ecosystem, it aims to empower startups, industries, and government agencies.

Data-Efficient Learning

A significant focus of BharatGen is on data-efficient learning, particularly for languages with limited digital representation. The initiative aims to create effective AI models that require minimal data, addressing the needs of underserved languages through collaborative research with academic institutions.

Future Roadmap

BharatGen’s roadmap extends to July 2026, encompassing milestones like extensive AI model development, experimentation, and the establishment of benchmarks suited to Indian requirements. The project also aims to facilitate the scaling of AI adoption across various industries and public initiatives.