How seasoned IAS officer Abhishek Singh is leading India's AI mission
A 1995-batch IAS officer with over two decades of experience, Singh has learnt the ins and outs of governance. Now, he is leading India's AI mission with a focus on foundational models, applications to leapfrog global competitors
Abhishek Singh, Additional Secretary, Ministry of Electronics and Information Technology, and CEO, IndiaAI Mission
Image: Amit Verma
Abhishek Singh’s career trajectory has taken him down some unexpected paths, like building roads, schools and hospitals, and even managing crisis situations like droughts and communal riots. A 1995-batch IAS officer with over two decades of experience, Singh has learnt the ins and outs of governance. Now, he is leading India’s AI mission.
“I’ve been working on AI initiatives since 2018, and the AI strategy got approved in 2019,” Singh says. Fast forward to 2024, the IndiaAI Mission was launched, and Singh’s work took on a new dimension. “It’s been a long journey, but we’re finally seeing the pieces come together,” he adds.
Singh has been a part of Digital India and collaborated on securing approvals for the Global Partnership on Artificial Intelligence (GPAI) chair. India was chosen to take over as the council chair in November 2022. The IIT-Kanpur alumnus is currently posted as additional secretary, Ministry of Electronics and Information Technology, and chief executive officer of the IndiaAI mission.
India’s national strategy for AI aims to establish the country as a strong AI workforce and leverage AI for public services, value creation, startup support and job creation, explains Singh. “To achieve this, we formed working groups with academia, researchers, startups and entrepreneurs to identify gaps.” These gaps include compute capacity, lack of Indian foundation models, need for dataset platforms, support for skilling and startups, development of applications, and tools for safe and trusted AI.
Such findings led to the development of a seven-pillar approach through a consultative process with key stakeholders. The plan is a whole-of-government initiative, recognising that AI cuts across domains like health care, education, logistics and agriculture. “Our strategy aims to build AI applications and models across the ecosystem, fostering a collaborative approach to AI development and adoption,” he adds.
India is on track to build its own foundational models. The country has taken a big step by funding Sarvam AI and plans to support 10 to 12 more startups in this area. “Our efforts span the entire AI stack,” Singh explains, from developing applications—30 are in the works—to building a datasets platform. The focus is on empowering Indian companies to play a key role in the AI ecosystem.
While India may be behind China and the US in some areas, Singh believes “our skills and capabilities can help us catch up and potentially leapfrog ahead, as we’ve demonstrated in cutting-edge research areas like space, missiles and electronics manufacturing”.
Many startups in India are working on foundational models, “and we expect these models to be developed and deployed by the end of this year or March next year”, says Singh. One of the key focus areas is on developing a voice-based large language model (LLM) that will enable users to interact with the model using voice commands and receive outputs in voice format. This endeavour is particularly important for India, given its linguistic diversity, with 22 official languages and thousands of dialects. The voice-based LLM will be trained on Indian datasets, allowing users to query the model in any language or dialect and receive responses in their preferred language.
The IndiaAI Mission has an investment of around $1.2 billion. Additionally, initiatives by the Ministry of Education, Centres of Excellence and Department of Science and Technology will contribute to an overall investment of $3-5 billion across departments and state governments.
According to Singh, the biggest hurdles to AI development were compute availability, datasets and funding. “We’ve made progress on all three fronts,” he says. The team has empanelled 14,000 GPUs in the first round and secured 18,000 more in the second. By year-end, India expects to have around 50,000 GPUs available, which should largely solve the compute issue. The newly launched AI Kosha platform provides access to datasets, and the IndiaAI Mission is funding foundation models and startups.
The potential applications of AI in India are vast, from health care and education to agriculture and climate change. Singh mentions that 30 AI applications have been selected and are being shared with relevant ministries for adoption. “We’re working closely with stakeholders to scale up and deploy these applications at a population level,” he adds.
On AI regulation, Singh says, “We’ve released a governance framework and are developing tools for AI regulation, including bias mitigation and deepfake detection.” India believes existing laws are sufficient to regulate responsible AI development but may reassess if needed.
“The AI mission is working well, and we aim to remain among the top five AI nations in terms of capability, use and talent in the world,” Union Minister Ashwini Vaishnaw tells Forbes India.
Singh emphasises the strategy is to promote Indian products that can stand alongside global ones—not to ban international products. He says Indian products, trained on Indian datasets, will offer better responses and tailored solutions, ultimately gaining popularity among users. The focus is on fostering innovation and competition, allowing the best products to emerge, he concludes.