NASSCOM President Debjani Ghosh spoke to DIGITIMES Asia about India's AI future.
After receiving a green light for a US$1.25 billion investment for its five-year AI mission, India's tech industry is preparing to deploy AI across the country.
At the forefront of this initiative is the National Association of Software and Service Companies (NASSCOM). NASSCOM is the largest non-governmental trade association in India's tech sector.
NASSCOM is dedicated to scaling up future-ready talent and fostering innovation across industry verticals. It aims to unlock new market opportunities for startups and corporations.
In an interview with DIGITIMES Asia (DT), NASSCOM President Debjani Ghosh (DG) talked about the association's role in upskilling talent for India's AI mission, nurturing unicorns in the AI era, and the potential of AI to address massive labor shortages.
DT: Why is it important for India to attempt to build its own AI ecosystem now through the recent approval of the US$1.25 billion India AI mission, and what impact do you foresee it has on the larger trajectory of the country?
DG: So the AI mission is set up not just to ensure India can develop their sovereign AI. But India can also deploy artificial intelligence in core sectors, and core industries. To get that benefit of a productivity boost, which is the promise of artificial intelligence we need to ensure that we unlock that accelerated growth potential because the growth will boost both the labor and their productivity. India's labor is young. But we need to ensure that productivity grows even faster than what is growing today, to help the country reach the goal of becoming a developed nation of 30 trillion GDP by 2047.
DT: NASSCOM is the not-for-profit industry voice for the over US$250 billion technology industry in India. In terms of the India AI mission, how does NASSCOM envision its role in this mission?
DG: At NASSCOM, we focuses a lot on the people part of it. We are the scaling partners for the government's frontier technology, we have an initiative jointly with the government, which is called Future Skills Prime, which is developed and designed to work with industry to get industry to come in, collaborate with us, and help us build out a talent surplus in India for digital skills.
Ghosh sees the demographics of AI talents in India in terms of a three-tier talent pyramid.
At the top of the pyramid are the world-class AI professionals and engineering talent. "A million or two out of the eight million engineers that we have in India are world-class AI developers," she said.
The middle of the pyramid is made up of white-collar workers like journalists, doctors, lawyers, and engineers, Ghosh said. They have the skills to operate AI tools, she added.
She said NASSCOM focuses most on the white-collar tier. The organization hopes to help people in that section of society gain the productivity boost that AI promises to offer.
The bottom of the pyramid is where the rest of society sits, Ghost explained. The government will focus efforts to instill digital, or AI literacy in this portion of society Ghost explained.
The goal is to teach people to use the basic tools for their convenience in a safe manner, Ghost said.
DT: India has a thriving unicorn scene especially when it comes to tech. What sort of role and assistance does NASSCOM provide to these tech start-ups?
DG: India has around 31,000 tech startups. And when I say tech startups, I don't mean startups that have apps or are you know, using technology, but startups that are building technology, products, services, whatever have you. So we have roughly around 31,000 tech startups, out of which around 3000, are deep tech. NASSCOM's focus is really on the deep tech part of it. We must ensure that number goes from 3000 to a minimum of 10,000 by the end of this decade. And for that, you need to bring in patient capital, you need to provide market access, you need to provide mentorship. So we have a complete end-to-end initiative where we are working with the startups from mentoring them to connecting them with industry, connecting them with government, and even connecting them with global markets and with investors. So we have a pool of around 11 investors who have come together and created a large deep tech fund that is available for startups that are working in this space. Funding is the biggest challenge when you're when you're in deep tech.
DT: What do Indian startups offer in this regard in the grand scheme of India's AI mission?
DG: While currently a lot of models are being built, and invested in India, and do focus on Indian languages, many of them don't know enough about our culture. This is why they can never adequately capture our culture. So when it comes to building indigenous models in India, I think that's where startups and academia will play a very, very key role. They will have the creativity and the expertise to come and do it. Last year was all about Large Language Models(LLMs), this year, we are going to see a lot more focus on vertical agents, it's about building on top of LLMs vertical models that will solve specific problems like you will build a model for insurance claim verification, or for software engineering, etc. And I think startups will again play an important part going forward. We are working with over 100 genAI startups in India. They are all playing in that space of building very strong vertical agents in business and data, healthcare, education, entertainment, and manufacturing process automation.
DT: Building on that, more precisely how does NASSCOM assist startups and smaller companies in accessing new market opportunities regarding AI or otherwise?
DG: First of all, what we do is we bring them the data. The insights on what new opportunities are out there. NASSCOM has always looked out to see what's next. And that's one of our jobs. In terms of AI, we started more than two years ago, when we started talking to the industry, especially the large services sector. Working with McKinsey to come out with a report, in which we talked about what would this upcoming AI wave would look like.
This February, we did a refresh of that. We went back to them to talk about the new usage cases. For startups, it's at a different level. For example, just last week, we took around a dozen of Indian startups to the Middle East, because they don't have a lot of homegrown startups. So we got them to sit down with the large companies with the government to showcase their capabilities, it's like matchmaking. Those start-ups then come up with solutions for these entities.
One of the largest tire manufacturers in India, CEAT, wanted us to figure out the best and most efficient way of conducting fault detection for their products. Detection takes a lot of time. So they wanted to know how they could speed it up without losing accuracy. So for this request, we worked with quite a few startups to come out with solutions, and one of the startups was Radom Tech. Their solution was accepted by CEAT, and they are now being used and deployed across eight factory lines. They've significantly brought down the time and cost of fault detection.
Similarly, there are so many case studies where we worked with hospitals, paint companies, and cement companies. NASSCOM's co-creation program and center of excellence works with these large companies to understand their problems. We take the problems to the startups. We say come up with the solutions. And if three of them come up with a solution, we get all three to pitch to the company. And then the company will choose whoever they want for the business.
DT: Where do you foresee AI as a whole going this year? How will it evolve and what should we look out for in 2024 and beyond?
DG: I hope that we have left the hype of AI in 2023. I don't think we will completely. Some level of hype will linger. But I do believe that this year, we must get much more serious about the deployment of AI. It can't just be talk. We see the reason all of us got so excited about AI. I think it was Goldman Sachs that first came out with a report saying you're going to get a 7% boost in productivity. Now, why is that so important? Think of countries like us, Taiwan, and Japan, with an aging population where your population is dropping out of the workforce because of age. Automatically, your growth gets impacted, because growth is a function of labor and productivity.
Now, the only way for you to get the growth back on track to pre-COVID days, under the circumstances of not being able to boost labor because of your aging population, is to go after productivity. And again with the promise of AI, we can accelerate or boost productivity. And I think we haven't yet fully been able to unlock that potential of AI to improve productivity across core sectors. To do that, you must deploy AI at scale across sectors. It hasn't happened yet. There have been a lot of good POCs and a lot of good pilots across the world. But you haven't seen that scale happen. I think it's time to get serious about scale. So for me, that's the first step.
Right now we're spending a lot of time on regulation, government after government. But what worries me looking at regulation through a lens of existential risk. Something bad could happen. But I honestly don't believe that's the right lens to apply. I think we cannot approach regulation from the lens of fear. We must approach regulation from the lens of opportunity because that's what the world needs now. We have to ensure that as we unlock artificial intelligence across sectors, and with the right oversight.
I'm not speaking for NASSCOM here, but personally speaking. I think this is where the world really can learn from India's success. India has had a tremendous amount of success in building and scaling technology to solve core problems like financial inclusion, and health care. And we have done this with the building of digital public infrastructure in India. We have built our digital public infrastructures (DPI), IT sector by sector, we've built it in finance, we've built it in health for vaccination. Our finance DPI, which is an additional public infrastructure, monthly transactions have crossed US$9 million. It's more than all the credit cards put together. Vaccine, The DPI for vaccination has over a billion vaccines delivered.
All these instances have shown with the right usage when you're solving the right problems, you can achieve the population-level scale of technology. But to do that you must build trusted technology. And to build trust in technology, you need an open ecosystem-led initiative, and not a few companies doing all the work.
But you need to make it open for the entire ecosystem to come and play. And that was the secret sauce of the interoperable, open, DPI ecosystem. And I think there are some very important lessons for us to learn when it comes to how we deploy AI at scale.
