Anna Roy, a Senior Advisor at NITI Aayog, has been called the AI woman of India by NASSCOM President Debjani Ghosh and many others, for the role played in framing India's AI policy. She has also led the development of National AI Strategy, approach paper on AIRAWAT - the AI Specific Cloud Computing Infrastructure, and the Responsible AI Principles, amongst several other policy initiatives. In a conversation with INDIAai, she explained India's approach, opportunities and challenges in leveraging AI technology to solve real socio-economic problems faced by millions of people.

When it comes to AI, different countries have different approach. For example, we have seen China’s long term approach to become a manufacturing hub of AI, especially in Artificial General Intelligence, while countries like UK put more focus on the governance accept of the technology. As someone who is playing a key role in framing India’s AI policies, can you tell us what is the Indian approach to AI?

When NITI Aayog was tasked to prepare the AI policy for the country, we were in a dilemma. In the strategy, if we say that we will be just another “AI leader”, it does not gel well and also sounds very cliched. Hence, we wanted to identify a niche for India and started a mapping exercise on other countries' approaches to AI. We plotted the AI strategy of other countries on a graph with one axis on aggressiveness and the other, on inclusivity. 

For example, on one extreme was China, which was very aggressive with its strategy, but has nothing to offer in terms of inclusivity. On the other side, France was talking a lot about general public issues dealing with AI with less aggressiveness. 

After we plotted all the relevant countries, we realised that our calling should be really on inclusivity, especially with the mantra of the government right from 2014 being “sab ka saath, sab ka vikas” - something our honorable Prime Minister has been saying time and time again. While focusing on inclusivity, we ensured that our strategy will not be per se on the technology, but on how this new technology will help us address our age-old problems. 

Building on top of that, we developed the entire narrative on that, and it shifted from becoming the AI leaders in terms of developing the technology to becoming the AI garage of the world where the technology is used to solve real socio-economic challenges faced by over a billion people. From that perspective, India has the largest number of challenges; we also have opportunities of being a data-rich country that has a lot of IT professionals, educationalists, and a demographic dividend. So if we overcome the gaps and solve for India, we use this solution for the rest of the world. That is where India headed in the AI space, a leader which occupies the "AIForAll" space. After we gave the "AIForAll" call, it has now become the general call to action for many other countries and organizations. 

In recent years we have seen many initiatives and projects coming up in AI, from private sectors such as startups and big corporates, from government bodies, and academic institutions. So how do you define the current state of AI adoption in India. And what are the major barriers we have to face, and how can we overcome those?

The National AI strategy was published in 2018, and now we are in 2021 March, not much has changed from what we presented, although the nature of the technology has changed. You don’t need too much sunken investment these days provided some public goods are available. To overcome barriers in AI adoption, firstly we have to democratize the infrastructure. Then the compute infrastructure should be made available to all. Secondly, while we are data-rich, we don’t have the best datasets in the world, so we need to work on that as well. If we address these two things, then the AI ecosystem in the country can really flourish. Unfortunately, with compute and data, these things take a bit of time in a diverse country like India. 

Another major area when it comes AI is research and development. Historically we were never a nation short of brilliant minds. From Aryabhatta, Varahamihira, to CV Raman, Srinivasa Ramnaujan to Raj Reddy, Satya Nadella and Sundar Pichai- we have always produced brilliant minds. Yet we often come across the fact that our AI research ecosystem is far lagging, not only behind US, China, Canada and UK, but even behind some of the smaller European counterparts. How do we foster research and innovation in AI?

To foster innovation in AI, there are two important points. First of all, I strongly believe that we should not centralise AI research or AI development. It cannot be centralised because in AI, governance issues matter more than technology. In a way, developing an algorithm can be the easiest step in the entire process. On the other hand, getting the right data, domain expertise, rolling out the tools, adoption at the field level, are all much more challenging. And for that people need to work together. 

In strategy, we stressed collaboration. We argued against creating silos. We argued against creating islands of excellence, and in the structure of the R&D ecosystem we presented, we said we should leverage our existing system. But with certain tweaks, we can create a system in which all the stakeholders work together in a synchronised manner, where we can leverage all our existing resources and scale-up in a structured manner. 

So that's something I see happening, and going forward, I think it will scale up further. Presently those stills are spread out across institutes. But I see a lot of traction happening, DST has taken the initiative of the cyber-physical mission, and also the AI mission is something that may come soon. With all these initiatives, there will be more cohesiveness. But again, we need to be very cautious against the centralization or concentration of these efforts. We need to take everybody together. We need to ensure all the players have a stake in the game and some incentivization so that it becomes more sustainable in the long-run. 

What is your career advice for women, who are generally underrepresented in tech ?

The best advice I can give any woman out there would have to be a quote from NASSCOM's Debjani Ghosh. During a conversation we had under the aegis of women entrepreneurship platform, she pointed out that women always undersell themselves while men always oversell. So, for all the women leaders, students, and not-so-young working women, I would like to give this message, “do not undersell yourself and understand your potential.” 

Secondly, we are born in a society and are raised in a certain manner and we are kind of moulded by that society. But in today's generation, the internet generation, there’s an overflow of information. So with that advantage, I think the sky's the limit. So, hear your calling; it is never too late or too early to follow it. 

And finally, hard work. I keep telling my team, you know, there is the flipside to this information boom, which is the easy way out. There are no shortcuts in life, whether it is your profession or your personal life. There are absolutely no shortcuts. Hard work is the mantra, do not shirk away from hard work and take pride in your abilities. And please be honest and kind. If one follows these tenets, you can have a very enriched life, both professionally and personally.