Add thelocalreport.in As A Trusted Source
Abhishek Singh, Director General (NIC), Additional Secretary, MeitY and CEO, IndiaAI Mission, is not trying to sell a vision. He’s trying to make sure it works. In a rare unscripted conversation at the grand finale of CNBC-TV18 HealthX Elevate presented by Optum India, he explained what the government is enabling, where it is lagging behind, and how trust – not just technology – will ultimately define India’s healthcare AI story.
The conversation, moderated by CNBC-TV18’s Ashmeet Kumar, discussed infrastructure, regulation, digital health innovation and the growing impact of Global Competence Centers (GCCs). And he did not shy away from contradictions. India has the talent, data and ambition. But it also has shortcomings in security measures, public trust, and real-world deployment.
New Diagnostic Partner
The session started with an anecdote: A senior executive has a health concern, gets a CT scan, a Bengaluru startup analyzes the scan with AI, and finds that the initial readings missed something. It was a useful starting point, because it asked the question many in the room were already wondering: Is this the future?
Singh did not defend himself. “In some cases, such as tuberculosis and some diseases, AI-generated results may be better than some radiologists,” he said. Not all – but some. He said the difference is in the dataset. While a radiologist brings his or her personal experience, an AI model trained on thousands of cases across institutions brings something else: cumulative expertise.
This has already started to be seen in practice. India’s health regulators – including ICMR and CDSCO – have approved AI models for use in diagnostic protocols. And the logic is straightforward: if AI can assist, enhance, or even outperform in limited contexts, then it should be used – provided it is tested, proven, and implemented with care.
Public Sector Weight, Private Sector Speed
When he talks about AI in healthcare, Singh speaks like someone who has seen systems break down under volume. Be it AI-powered solutions that help public hospitals better manage OPD overload, or ICUs in remote areas that operate entirely on telemedicine protocols, he sees AI less as an upgrade and more as a way to keep the system from collapsing under its own weight.
This is what makes India an unusual test case. On the one hand, you have high-end startup solutions for speech-to-text clinical notes, multilingual doctor-patient translation, and autonomous radiology screening. On the other hand, you still have patients dying from untreated diarrhea because there are no doctors or information within reach. That duality is where Singh sees value – tools like NLP-based voice bots that handle healthcare queries in regional languages, or AI triage systems that help recommend the right tests to non-physicians.
But the expansion of those systems will not be driven by the government budget. Instead, he sees the government playing an enabling role. “We had a lot of discussions across the industry,” he said, “and the response was clear. We need to invest more in computation, in R&D, in building datasets and supporting foundation models. These are the objectives of the IndiaAI mission.”
The numbers are impressive. The government has promised ₹10,000 crore for this initiative. But private players have already invested equivalent to ₹20,000 crore to bring in 38,000 GPUs. Singh was clear about the intention: By subsidizing access, the government enables private investment that exceeds its own spending, and multiplies.
“Private players invested ₹20,000 crore to bring in those GPUs. We are subsidizing the access – so startups don’t waste all their capital trying to train just one model.” It is a strategy designed to lower the barrier to entry without centralizing control. And it is working. But Singh was quick to point out that expanding infrastructure does not mean ignoring the risks.
Trust as an Adoption Supporter
One of the sharpest analogies offered by Singh came when the conversation turned to public adoption. Asked whether Indians are ready to trust digital health tools, Singh talked about UPI. “When a vegetable vendor hears that ₹20 has been paid, sees the tick mark, he doesn’t even check. That’s how much trust there is in UPI. And it’s built at the bottom of the pyramid,” he said. “If we can do that in payments, we can do that in healthcare. But we have to earn that trust.”
And that trust, Singh made clear, cannot be separated from regulation. He does not argue against AI, but instead emphasizes responsible implementation. Singh argued that like a new drug or vaccine, AI models in healthcare should be deployed only after proper testing and regulatory approval.
That framework no longer fully exists today, Singh suggested. While the RBI has already released the draft of a responsible AI framework for financial services, healthcare still has no counterpart. “This needs to change,” he said.
New discipline around data
No conversation about health care data is complete without addressing privacy. Singh acknowledged that the Digital Personal Data Protection (DPDP) Act is an important step in the right direction. “We are waiting for the rules like everyone else,” he said. “The sign is – it could come any day.”
Meanwhile, Singh explained how the Act will apply to AI models: not sharing personal health data without consent; No processing beyond the purpose for which it was given; No retention beyond the permitted period. In other words, the same rules that govern human actors will apply to algorithms – once the framework is activated.
From India to the Global South – and beyond
When we think in terms of designing AI systems, India’s healthcare system doesn’t just have limitations – it has contextual constraints. A multilingual population. Shortage of doctors. Poor infrastructure. Many people do not have access to doctors, diagnoses or basic information. And a cost structure that defies most Western technology.
These are the design specifications for the future of health care innovation – not just in India, but across the Global South. If India can build systems that work here – multilingual, infrastructure-light, physician-agnostic, and affordable at scale – they won’t just be tailored for India. They will be world-ready.
That is why Singh laid emphasis not only on engineering capability but also on ownership. “Our engineers have contributed to the growth and expansion of almost every big tech company across the world,” he said, “which is why it is our priority to provide them any support to build products in India, rather than leaving India and contributing elsewhere.” He expressed confidence in India’s ability to create the next wave of AI products, “made in India, and will soon become global commodities in the coming days.”
That’s the bar, and that’s the opportunity. And when this happens, India will not only transform its healthcare system. It will reshape the meaning of healthcare everywhere.