Dr Rohini Srivathsa is a name that needs no introduction in India’s tech space. As the National Technology Officer at Microsoft India, she is one of the prominent thought leaders with an illustrious career that span across numerous reputed organisations in India and across the globe. An alumnus of the prestigious Wharton Business School, Rohini shared her in-depth insights and opinion on the role Artificial Intelligence in the new decade, especially from a post-COVID economic recovery perspective, and how we can channel the power of technology for social and economic transformation with INDIAai.

As one of the most important tech company in history, what is Microsoft’s Approach to AI?

When it comes to AI, Microsoft’s approach is built on three pillars. They are meaningful innovation, empowerment of people, and responsibility. Microsoft’s AI research began over two decades ago, and it was only in the last few years that we have seen AI innovation picking up an exponential pace. Thanks to the rise of data and computing power. 

Getting these innovations out in the hands of the people is one of the core tenants of Microsoft's AI approach. Today, from modern workplace to business applications and cloud with Azure, we are looking to bring more artificial intelligence embedded products and services that can empower people. 

Coming to the third pillar, responsibility, we started out by looking at the principles or ethical issues that would arise as AI become more mainstream. We have come a long way in that over the last three-four years and have reflected deeply on how to put these principles into practice at scale, not just for ourselves as we build our own product, but to help our customers and partners think about responsibility in the age of AI.

The Responsible AI principles that Microsoft have put forth are six, and we think of them as reasonably comprehensive, mutually exclusive, and very powerful. They are fairness, safety and reliability, security and privacy, inclusion, accountability, and transparency. These are principles to contextualise every single AI use-case, especially in the context of a particular customer, situation, problem statements, or the stakeholders involved. 

It is a very dynamic way of thinking, and in our experience, we've come to see that it's important to take things from principles into practice. And these practices can be taken be in the shape of guidelines or checklists in very specific use-cases or technologies. They take into consideration what might be the governance methodologies or approaches organisations might take. It is a journey from principle to practise, and there is quite a bit of organisational maturity and know-how we have developed over the period of time on this. 

As the world continues to go through one of the most challenging times of this century, the role technologies play to overcome these difficult circumstances is becoming crucial. How do you think COVID19 impacted the technological landscape, especially concerning artificial intelligence? 

In the context of COVID pandemic and what it is doing to the economy and the world at large, the concept of tech-intensity is becoming more relevant. Tech-intensity is the combination of adoption of latest technology interspersed with internal capabilities that are built to leverage that technology. This is important because just adopting technologies from the outside is not good enough going forward, rather the important thing is your ability to quickly utilise what's out there and build on top of it. In essence, that ability to build internal capabilities in your own secret sauce, and then adding the technology on top of it is becoming more important now.

As Satya Nadella, the CEO of Microsoft pointed out last year, the pandemic drove digitalization worth 2 years happen in just two months.  This means, every organization, whether traditionally digital or not is becoming more and more intense from a tech perspective. 

The tech-intensity is now increasing across sectors, especially in the case of our COVID-19 response. And in recent months, we are seeing a rebound phase, where the economy is coming back up. As a result, businesses across sectors are turning their attention on reinventing themselves as the shadow of the uncertainty regarding how their industry will look like in the post-COVID world looms large, forcing them to resort to AI for reimagining their businesses. Therefore, a lot of AI is getting built into how people are thinking about reimagining business. 

You mentioned how businesses are trying to reinvent themselves as a result of COVID. As a result of this new mindset, in the post-COVID economic scenario, which all sectors will be disrupted by AI?

Going forward every domain is going to see a split between the physical and the cyber world. Every industry is going to wrestle with that cyber-physical play, which will play out differently across industries.

For example, the way healthcare has been delivered in the past, assuming a physical presence, is getting disrupted. Especially with the advent of telemedicine and the ability to connect specialists to parts of the country where the supply of high-end expertise is not available. 

Now, if you build on top of that capability, suddenly you can potentially have a virtual assistant that can do initial triages for patients and offer personalized initial assistance, that can then be supplemented by an expert. It opens up a lot of opportunities for data and AI, and it will even help bridge the gap in access to quality healthcare. This process will play across the sectors and will accelerate AI disruption in industries that are already getting more and more digital, like financial services, telecom, and to some extent, retail.

The industries which were not feeling the high intensity or urgency to become digital previously, like manufacturing, are now getting pushed to look at how much of the business or the operations need to remain in the physical world and how much of it can become in the digital world, as a result of the pandemic.

Coming to India, in the last couple of years, we have seen numerous AI initiatives from the public sector and private sector to promote AI adoption across the country. How can we fastback our AI ambitions, and what are the challenges we need to overcome?

In some ways, India was ahead of the curve or fairly progressive when the National AI Strategy was rolled out almost more than two years ago. The National AI Strategy, which as brought out by NITI Aayog, is a very well laid out initiative. However, the question we have to wrestle with is how can we make that strategy work for the country to really unlock the value from data? Because beyond strategies, it is also about a set of actions that the country needs to take. Actions in terms of the right datasets, the right kind of infrastructure and tools, the right commitment to investing in AI, and in skilling. 

We had done a survey with IDC where we actually found that 33% of organizations in India today have started on their AI journey, but the remaining are yet to really think of AI as an important competitive lever. That is a concern because AI is one of those technologies where you need to start with experimental trials while building a strategic view to it. 

The survey has also found what are the top challenges for organizations in India. The most important one was that AI is not yet being seen as a strategic priority. Unless that happens, the leadership commitment to investing in AI, will not happen at any organisation. The second challenge is that there is a low level of thought leadership and leadership commitment to investing in AI. The other issues include not having enough investment as well as scale-up of infrastructure and tools, and then of course skilling becomes important for making all of them work.

We have seen so many incredible AI initiatives coming from Microsoft in recent years, especially here in India. Can you tell me some of the interesting projects Microsoft AI Research in India has currently undertaken?

Microsoft Research globally is nearly 30 years old now, and we've got centres around the world. Microsoft India is about 15 years old and works across multiple areas, from AI or image recognition to human-computer interaction.

There are some of the interesting work that Microsoft India has been going on AI.  One is a device that can be used to monitor the state of a driver and how the vehicle is being driven in the context of a road environment that the vehicle is in. The algorithm can look at the driver's methodology and can give some coaching if required.  It is called HAMS or Harnessing AutoMobiles for Safety. It's a low-cost sensing device that can be employed in a regular smartphone and can be it can be used by a Department of Transportation, as well as by drivers, or even the likes of Uber and Ola.

Microsoft India Research’s approach is to push the boundary of knowledge in new areas and to work on technologies which are relevant to economies or even demography like India, where we look into using very low-cost sensors and such. Some of the really path-breaking work is happening on technologies which are relevant to the emerging markets.

Furthermore, Microsoft Research works very strongly on creating tools for societal impact. There is an MSR India Centre for Societal Impact which works with a range of underserved communities whether it is academia, startups or NGOs on solving social problems using cloud and AI. The centre also focuses on social entrepreneurship. 

Speaking of tools for societal impact, what are AI initiatives from Microsoft that focuses on social empowerment?

There is a whole range of areas that Microsoft is working on, and many of these are global programs as we can leverage a lot of work across geographies and learn from each other. One of them is the AI for Health initiative, a five-year commitment to work with non-profits, researchers and educational institutions on healthcare issues, especially on the COVID-19. Many of the open data initiatives have come from this, like making COVID related literature available in a searchable method.

Accessibility is another area that Microsoft, starting from Satya Nadella, really feels very strongly about. We have been working on tools that use AI to provide accessibility to millions of people with disabilities, such as the AI-powered for visually impaired.  

Another area is AI for Humanitarian Action. There is a lot of work happening on helping refugees, disaster recovery, for stopping child trafficking and aiding people who have been displaced because of climate change or human rights issues. Here we are working with the likes of the United Nations and other humanitarian agencies around the world. And this a 40-million-dollar five-year program, as we are looking to leverage AI for humanitarian issues. 

AI for Earth focuses on sustainability across areas of Agriculture, Water, biodiversity, climate change. Another interesting one is AI for saving Cultural Heritage. We don't realize that many of the pictures or artefacts or sculptures or various cultural remnants can be both preserved and enhanced digitally. We can also piece together components that are sitting in museums around the world using AI and gain a lot of interesting insights, especially from ancient scripts. 

When we speak about social challenges, gender inequality is one the serious ones, and now we are seeing numerous cases of gender bias in AI algorithms, which can only be prevented with more women participating in the process of creation of algorithms. As a champion for gender equality, what are your thoughts on how can we bring more women into technology, especially in terms of AI?  

There is a huge gap in the technology world between the demand and supply of the right kind of talent. We all know that the supply of skilled talent is not up to the demand, even globally speaking. Therefore women participation, just from a very fundamental demand-supply perspective is important. 

But that's a very first- level thing. The second important thing, in an Indian context, is that the participation of women in the Indian economy is one of the most untapped potentials. If you look at the rise of the US in the 60s and their economic growth, it was very strongly correlated and coincided with the rise in women's participation in the workforce.

Importantly, AI also needs women, because the ability to think about a problem from multiple dimensions, not just from an analytical dimension but also from, the point of empathy, collaboration, etc, is important.  That is the cognitive strength that women naturally bring to the table. 

Furthermore, there are many studies which show that women leadership style and women organizational styles are different and therefore gender diversity in the workforce is only going to increase the effectiveness and an innovation capability of an organization.

Now coming to the question of how do we accomplish more women participation? 

I think in India we do have a lot of women entering STEM education and I don't think India has got that bigger of a problem of women not being interested in science and technology. I think what is important is their ability to see role models, the ability to picture their career as a long-term marathon and stay the course. Especially when they go through important life stages such as having a family or taking care of others, they should be able to come back into careers after a break. It is important to have mentors and the support system that helps them to do that. 

Many of those things are an ecosystem problem. It's not about getting more women but making sure that they are staying the course right. And therefore, factors such as the ability for women to upskill, being able to stay the course even if they take a few off-ramps becomes critical. Another important challenge is overcoming unconscious bias in that exists in organisations. 

So, women participation is not just about skilling, even though it is one of the aspects, but the major steps we should focus are in providing role models, encouraging long terms thinking in terms of career plans, empower them with the ability to on-ramps and off-ramps, and the issue of unconscious biases which is across all genders. And many organizations like Microsoft, IBM and others have been helping women coming back into a career. Those are good but we have to do more of these.