AI Summit | India AI Mission to focus on applications, datasets, compute infrastructure in next 12 months: CEO

India’s AI ambitions are moving beyond adoption toward intellectual property creation.

• India to focus on practical AI tools and home-grown base models
• Plans include expanding data, compute power, and AI adoption
• Ethical safeguards include deepfake guidelines and watermarking

Did our AI summary help?

India’s artificial intelligence strategy will concentrate on building practical AI-based tools, developing a large home-grown base AI model, expanding data resources, and increasing the high-end computing power required to run such systems over the next year, according to Abhishek Singh, Chief Executive Officer of the India AI Mission. The strategy is aimed towards strengthening the building blocks that allow AI technologies to be created and widely used across sectors.

Speaking to Moneycontrol on the sidelines of the India AI Impact Summit in New Delhi on February 17, Singh outlined a multi-pronged roadmap that places equal emphasis on technological capacity and adoption. “Priorities are getting applications rolled out, getting our foundation model established, expanding the databases and the datasets that we have, and expanding the compute infrastructure that we have,” Singh said, describing the operational focus for the coming 12 months.

Beyond model development and infrastructure creation, Singh indicated that enabling usage will be a central pillar of policy design. “We are building an influencing layer so that people are actually able to take advantage of AI applications and use cases,” he said, suggesting that the Mission’s next phase will address barriers to adoption across industries and public services.

Foundation models and AI infrastructure

Singh emphasised on the importance of compute capacity, a critical constraint in AI development. Expanding compute infrastructure typically involves investments in high-performance data centres, specialised processors, and cloud-based resources required for training and deploying advanced AI systems. Policymakers and industry participants have repeatedly flagged compute access as a bottleneck for startups, researchers, and domestic developers.

Robotics under focus

Singh framed India’s AI ambitions as moving beyond adoption toward intellectual property creation. “We are a tech talent nation – the whole world knows of that,” he said, arguing that India’s skilled workforce can support the development of core AI technologies as well as applied systems.

He identified a wide spectrum of opportunity areas. “Wherever it is developing foundation models or building robotics, physical AI for robotics, building the brain for robotics, using AI in sectors like agriculture, healthcare, education, manufacturing, supply chain, logistics – all these areas are there in which we can build IP,” Singh said.

AI safeguards

Alongside expansion plans, Singh stressed that regulatory and ethical safeguards are already being embedded into India’s AI framework. “Yes, of course,” he said when asked about ethical protections, pointing to a series of policy measures introduced in recent months.

Deepfakes – AI-generated or manipulated audio-visual content – have emerged as a global policy concern, with implications for misinformation, fraud, and reputational harm.

“We came up with AI governance guidelines, which define our entire governance framework,” Singh said. “We also came out with our deepfakes guidelines, which require platforms to remove such content within a finite time,” he noted.

Additionally, Singh highlighted transparency measures designed to help users distinguish machine-generated material. “We also came out with watermarking and labelling of AI-generated content – so those guidelines are already there,” he said. An institutional oversight mechanism is also in place, he added. “There is an AI governance group, which further determines how we will ensure that while we enable AI development, we will also limit any misuse or abuse of AI.”