Chennai: Infosys Co-founder and Chairman Nandan Nilekani has committed a fresh multi-year grant to AI4Bharat, an open-source research initiative at IIT Madras that is building foundational AI models for Indian languages, marking his second significant funding contribution to the effort.
AI4Bharat, in collaboration with Nilekani’s EkStep Foundation, has been instrumental in digitising Indian languages, creating datasets and developing AI models tailored to the country’s linguistic diversity.
The initiative has emerged as a key pillar in India’s AI infrastructure, powering speech, translation and text-to-speech tools used by public digital platforms.
These models are at the core of India’s national language platform, Bhashini, under the IndiaAI Mission. They enable multilingual access to governance, healthcare, finance and more.
From the Supreme Court’s SUVAS system, translating judgments into regional languages to NPCI’s voice-based UPI transactions and agricultural chatbots like Kisan e-Mitra, AI4Bharat’s innovations are reshaping how digital services are delivered in local languages.
“Inclusion begins with access–and for Bharat, that means language,” said Nilekani, underlining the importance of building digital infrastructure that works for every Indian, regardless of the language they speak.
AI4Bharat has also released extensive language datasets covering all 22 constitutionally recognised Indian languages. These are freely available to developers and researchers via the AI4Bharat website and are integrated into AIKosh, the open AI repository, launched earlier this year.
Prof V Kamakoti, Director of IIT Madras, stressed the relevance of AI4Bharat’s work in preventing an AI divide and supporting the vision of “AI for All.” He said, “With the generous support of Nandan Nilekani, I hope AI4Bharat continues to contribute significantly toward this national objective.”
The previous grant in 2022 had established the Nilekani Centre at AI4Bharat. Over the last three years, the initiative has shown a scalable model for building open, population-scale language infrastructure by combining community participation with scientific excellence.
Discussion about this post