India’s approach to AI must be guided by the Gatishakti model

While India is emerging as a prominent hub for AI patent filings, most of these are currently contributed by MNCs. Prof Rajat Agrawal, Professor, Department of Management Studies, IIT Roorkee, talks about the major challenges to AI growth, the unique problems that it can address for India, and how the startup system can take the lead.  

Widely recognised as the most transformational technology of our present day and age, Artificial intelligence (AI) is increasingly taking centre stage in government policies and business strategies around the world. A PwC report on AI projects that the technology could contribute around US$ 15.7 trillion in additional GDP to the global economy by 2030 and provide a boost of up to 26% for local economies.

This has precipitated a race among major nations like US, EU, Japan and China to dominate the global AI ecosystem.

China has stated its ambitions to become an “AI superpower” by 2030. The EU is looking to take the lead on AI regulation, and the US is looking to maintain its edge in the face of rising competition. India is a very important player in the field of AI, with WIPO also recognising it as one of the most important emerging new target markets for patent filing in this area.

But who is winning? This actually depends on what metrics you choose for making a judgement or even presenting an opinion.

The first patent in the field of AI was filed in 1980 in Japan. Globally, out of the top twenty companies filing patents in the field, twelve are from Japan. The first two are from the US – IBM and Microsoft. But when you view it in terms of universities and academic institutions, China is the front runner. Out of the top twenty institutions, seventeen are Chinese. Now countries like France, Germany, Korea, UK are also participating in the race.

Coming back to India, one must note that while there is a good increase in the number of patent filings in AI from India, a large number of these patents are filed by multinational companies. Availability of companies like Microsoft in India is helping us to improve our IP portfolio also. But this also points to a challenge – we need to develop our indigenous capability to work in the field of AI.

Why are countries like China in a better position than India? One reason is that the industry-academia interface is much stronger. Then there are very few obstacles in collecting data in case of China, which is the backbone for AI-related innovation. So that creates a kind of an enabling environment for AI research for Chinese universities. Here in India, many ethical clearances required for collecting data and using it for research purposes.

Multiple government bodies or ministries are involved in AI-related activities, and they need to be consolidated. Take the example of the PM Gati Shakti Project. The Prime Minister has taken this initiative to put different ministries involved in infrastructure-related projects on one platform. I think with this type of initiative, the government is also realizing that we need single window solutions for many problems. This is being done so that efforts can be concentrated and we can certainly leverage our investments. Similar initiatives are required in AI, where multiple stakeholders are there.

New problem areas mean new opportunities

Addressing these issues is important for AI to flourish in India, not just as a business opportunity, but also as a technology to address some larger problems. Note that IBM and Microsoft are having the maximum number of innovations in the field of AI, with around 8000+ and 6,000+ patents respectively. Most of their patents are in the field of computer vision and natural language processing.

But countries like India have different types of needs, which also bring interesting new opportunities. The first that comes to mind is healthcare. Covid has been a wake-up call for India to strengthen its health care system. The Aarogya Setu app became a very important tool for tracking Covid patients. Now we need to see how India can use that database for preparing a more robust health care system for its citizens. Already some ground has been covered and we need to capitalize on that foundation. If we can provide meaningful health care to our population, someone who is very poor can also live happily.

At the same time, because we are in the knowledge economy and these days, education plays a very important role. So how can AI technologies help provide education to far off areas of our country? During Covid, we saw a rapid increase in various education technology platforms. At IIT Roorkee itself, we signed more than 10 MoUs with different edtech partners to provide online classes. IIT Madras already has this very popular example of National Programme on Technology Enhanced Learning (NPTEL). Now, various other ed-tech platforms are coming and we can actually democratize education, which was earlier considered to be reserved from some elite sections of the population.

The third important area where AI should play a role, particularly from India’s point of view, is agriculture. We are still known as an agrarian economy. AI inputs can be invaluable to farmers in areas like producing healthier crops, pest control, soil management and optimising the supply chain.

There is vast potential in India’s startup ecosystem to tackle these problems through AI, but unfortunately it does not get realised. The academia level projects that we see our students complete are one strong example. But in a campus like IIT, they get such good offers from corporations that they do those projects only as a passion or as per the requirements of the curriculum. Those opportunities are not further taken up into real start-ups.

So we need to promote entrepreneurship – the knowledge for creating good start-ups – from the campus itself. That can be one baby step in the right direction. Secondly if a start-up is already there which is not so proficient with AI-based technologies, we can provide additional inputs to them. At the college level it will be much easier to do that. There are different groups that work in silos, and we need to integrate them to map societal problems to the most appropriate AI solutions.