A day after the Indian government’s Economic Survey 2026 suggested that India will priortise SLMs over LLMs, India’s Ministry of Electronics and Information Technology (MeitY) Secretary S Krishnan has called on the country to move beyond its emphasis on generative AI. Instead, the IT Secretary has urged the technology scene to prioritise the development and deployment of smaller, sector-specific AI models. He argued that such targeted AI applications could deliver greater productivity gains across key economic sectors like healthcare, education, manufacturing, and agriculture.
Speaking at the “Democratising AI Access through Distributed Compute: Perspectives from the Global South” event on January 30, Krishnan questioned India’s widespread fixation on generative AI.
“Why are we so obsessed with the generative AI part? Why are we not looking at other aspects of it?” he asked. He stated that previous generations of AI, when fine-tuned for specific tasks, often produce superior results in practical applications compared to broad generative systems.
IT Secretary warns about Gen AI obsession
Krishnan also highlighted that generative AI represents just one segment of the broader AI ecosystem. He urged India to avoid competing head-on in the race to build massive foundational models and instead position itself as the “application and use case capital of the world.”
“We believe that India has the potential to be the application and use case capital of the world,” he stated, encouraging startups and innovators to create tailored solutions that drive real-world impact across the economy.
India should build sector-specific AI to be productive
The IT Secretary stressed the transformative potential of smaller, specialised AI models in productive sectors. These models, he noted, can make a “significant difference” in areas critical to India’s growth and development. By focusing on sector-specific use cases rather than general-purpose generative tools, India can achieve more efficient and impactful outcomes, particularly in resource-constrained environments.
This approach aligns with India’s ongoing AI strategy, which rests on three core pillars – infrastructure, models, and data. Krishnan advocated for an open public-private partnership model to develop digital and AI infrastructure that serves not only India but also other countries in the Global South. On the data front, he pointed to the AI Kosh platform, which has already onboarded over 7,000 open datasets, with more being added continuously. However, he acknowledged challenges with siloed government data and called for greater openness, inviting private sector contributions. “Our job is far from being done,” he remarked.
Krishnan’s comments come at a time when India is pushing under the IndiaAI Mission to foster inclusive and frugal innovations in AI. The government’s efforts include supporting sovereign AI models, subsidising compute resources for startups and researchers, and promoting applications tailored to local needs, including regional languages and data protection.
