IIT Madras Proposes Responsible Frugal AI Framework for India By Elets News Network - 19 November 2026

Gajendra Circle fountain with two elephant statues on the IIT Madras campus in Chennai, surrounded by palms

The Indian Institute of Technology Madras (IIT Madras) has proposed a Frugal AI Framework for India, urging the country to build smaller, specialised and multilingual AI systems instead of chasing ever-larger frontier models. The white paper from its Centre for Responsible AI (CeRAI), announced on 19 November 2026, calls this approach “Responsible Frugal AI”.

The idea is simple: an AI system should meet clearly defined needs while using the least compute, energy and money possible, without giving up reliability, robustness or accountability.


What the Frugal AI Framework proposes

The paper, titled A Frugal AI Framework for India, describes a third path for the country. It sits between copying capital-intensive frontier-model programmes and remaining a downstream adopter of foreign AI.

Its main recommendations are:


  • Make frugality the default engineering principle for public-interest AI, with priority for right-sized, domain-specific and multilingual systems.
  • Invest public money in shared AI infrastructure and common platforms, or “public AI rails”, rather than isolated and duplicated systems.
  • Build modular, interoperable systems designed for India’s diverse population.
  • Deploy generative AI selectively and judge systems on practical utility, resource use, accountability and social value, not benchmark scores.

The approach echoes other recent sovereign AI efforts covered by Elets CIO, such as the IndiaAI Mission’s release of the Soket EKA-1 and ZenteiQ Dhir models.

Why IIT Madras says smaller AI is smarter

Prof. V. Kamakoti, Director of IIT Madras, released the paper at the Industry Conclave on AI Governance 2026, hosted by CeRAI at IIT Madras on 22 October 2026. “AI does not have to know everything to be useful. It has to know the right things, deeply, for the task at hand,” he said, adding that much of today’s AI uses enormous amounts of electricity and water for tasks a smaller, well-designed model can do as well.

Prof. Balaraman Ravindran, Head of CeRAI and the Wadhwani School of Data Science and AI (WSAI), said India does not need to replicate the AI development models of the US and China. It can instead build on its strengths in frugal innovation, engineering and digital public infrastructure.

Dr. Gokul S Krishnan, Research Director at CeRAI, said institutions should first establish whether AI is needed at all, then pick the smallest model or technique that does the job. Rules-based systems, retrieval-based approaches and conventional machine learning should be used wherever they are enough.

A seven-point checklist for AI adoption

For government departments and enterprises, the framework sets out a seven-point checklist. It asks institutions to:

  • Justify why AI is being adopted
  • Define operational constraints
  • Select the appropriate model
  • Test in real-world conditions
  • Track lifecycle costs and resource use
  • Ensure documentation, independent audits and human oversight
  • Keep manual alternatives in place

It also recommends that organisations publish deployment failures, so others can learn and accountability improves, according to a DT Next report.

Who wrote the white paper

The paper was written by Prof. Balaraman Ravindran, Venkataraghavan KS (Senior Advisor, CeRAI), Dr. Gokul S Krishnan, Prof. Balaji Srinivasan of WSAI and Prof. Balkrishna C. Rao of the Department of Engineering Design, IIT Madras.

It follows an earlier IIT Madras CeRAI whitepaper on AI’s impact on India’s workforce, and the institute’s work on applied public-sector AI such as the IIT Madras-built LLM 2.0 for CAG auditors.

What it means for CIOs

For Indian CIOs facing rising GPU and inference bills, the framework offers a practical test: right-size the model to the task, measure total cost and energy use, and keep humans in charge. It fits the wider push for responsible enterprise AI and the themes of AI governance and digital sovereignty discussed at the World AI Summit 2026 in Bengaluru.

Key takeaways

  • What: IIT Madras CeRAI white paper, “A Frugal AI Framework for India”.
  • Core idea: Responsible Frugal AI, meaning right-sized, specialised and multilingual AI with the least resources.
  • Released by: Prof. V. Kamakoti, Director, IIT Madras.
  • For adopters: a seven-point checklist covering need, model choice, testing, costs, audits, human oversight and manual fallbacks.

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!

Related artificial intelligence