IndiaFocal.

India, in focus.

National

Representative image · Photo: IndiaFocal
Representative image · Photo: IndiaFocal

BHASHINI CEO at IIM Calcutta Conclave: AI Must Adapt to India's Diversity

BHASHINI CEO Amitabh Nag addressed Lattice 2026 at IIM Calcutta, stressing voice-first multilingual AI and ecosystem design for population-scale impact.

The annual business conclave of the Indian Institute of Management Calcutta, Lattice 2026, was held in Kolkata under the theme 'The Capability Decade', drawing students, industry leaders and professionals for discussions on artificial intelligence, technology, innovation and leadership.

The conclave's Chief Guest was Amitabh Nag, Chief Executive Officer of the Digital India BHASHINI Division (DIBD) under the Digital India Corporation, Ministry of Electronics and Information Technology. He spoke about India's experience of building and deploying language technology as Digital Public Infrastructure, focusing on the distance between AI models and experimentation on one side and real-world adoption at scale on the other.

Addressing the challenge of reaching a diverse population, Nag described the digital, language and literacy divides as interconnected, and argued that technology must adapt to people and their contexts. He underlined the value of voice-first and multilingual tools in widening access, noting that India's diversity demands AI systems that account for differences in language, dialect, domain and context.

The session offered a live illustration of that approach: BHASHINI's Shrutlekh was deployed as Nag spoke, delivering real-time translation and transcription and showing that language AI can function in an unscripted setting rather than only in controlled experiments.

Nag also discussed the organisational and product decisions that determine whether emerging technologies succeed. Drawing on BHASHINI's work, he spoke of building with people and for people, including community contributions to language resources through BhashaDaan, and stressed clear use cases, rapid experimentation, learning from failure and short feedback loops. Bringing research and product teams closer to real-world requirements, he said, is central to turning ideas into deployable solutions.

Reflecting on the platform's journey, he said the hardest problems were almost never about the model, but about market design, incentive design, organisational design and ecosystem design. Diversity, he added, is the new normal and demands fast, quick adaptation.

The discussion extended to partnerships, open infrastructure, co-creation and teamwork, with Nag noting that organisations working on emerging technologies need a culture of conversation, experimentation, brainstorming and partnership, along with infrastructure that lets many participants contribute to a larger ecosystem. Other sessions at the conclave examined emerging technologies, business and leadership, reflecting a broader emphasis on capability building beyond technological development alone.

Closing his interaction with students, Nag urged them to look past individual applications and consider the larger platforms, standards and ecosystems they could help create, asking what ecosystem standard they can establish and how quickly others can be enlisted to build network effects around it.

According to DIBD, BHASHINI powers more than 800 government websites, has enabled over 9 billion cumulative AI inferences, processes more than 24 million AI inferences daily, and supports 36 Indian text languages, 23 Indian voice languages and 35 international languages.