Development via AI

The Union Budget 2018 has one remarkable aspect when it comes to innovation in high-tech areas. It has a proposal that the Niti Aayog will launch a tiol programme of artificial intelligence (AI) for driving the government’s efforts for tiol development. This is a big proposal. But first thing first. The layman – all those poorer lot and the farming class for which this budget seems to mean – is not introduced to the hi-fi area of AI. So some simple discussion on the very complicated subject of AI is imperative. As the me itself indicates, it is not tural intelligence – that is, the intelligence of the human mind as it perceives, reacts, displays a range of emotions and behavioural traits, and goes in for decision-making. AI is essentially about machines, especially computers, that can simulate the human mind. Say, for instance, you have a computer with you that performs your tasks, however complicated, by applying its own mind. This machine can think and act as you do. It has a decision-making power. It may perhaps exhibit emotions! So your task becomes that easy. Now the layman will dismiss it as complete impossibility. But this is what is precisely happening in the realm of advanced computer science where one can develop neural networks in order to create a machine system that replaces man by virtue of it being a mind machine. AI research is flourishing in many prestigious universities like MIT and Stanford, and this has led to its increasing use right from automation, economics (especially the probabilistic part of it), banking and fincial services, social engineering, planning, language processing, robotics, optimization problems, operations research, medical diagnosis (including psychotherapy!), customer service, art creation, to even video games. Now this must be astonishing for the layman, but it is a fact of life bettered by the use of AI. This brings us to the imperative of continued AI research despite questions raised by both scientists and philosophers about the moral or ethical aspects of thinking machines – and how dangerous they can turn out to be for the human race if they go out of control. In fact, acclaimed physicist Stephen Hawking – credited with giving us some rare glimpses of the deeper mysteries of the universe – has gone to the extent of saying: “The development of full artificial intelligence could spell the end of the human race. Once humans develop artificial intelligence, it can take off on its own and redesign itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete and would be superseded.” But this is a different story.

The question now is whether the ambitious Niti Aayog programme can achieve its goal in a country like India where AI research is yet to take off in the real sense as it has in advanced countries. The proposal is a right one; the Modi focus is right – that of maximizing the use of technology for tiol development in all key areas where development is lacking and hampering the surge of the country as it aspires to be a power of global reckoning. However, a reality check must be in order. How many of our institutions or research centres have produced quality AI research as to be widely applicable? There is no great news coming in. Therefore, the first priority should be on quality research, for which institutions like IITs and IISc-Bangalore need to be funded separately for high-quality AI research. Secondly, since AI applications are diverse as already stated, experts from diverse disciplines need to be roped in so that the research becomes an inter-discipliry domain. For instance, one can have a mix of computer scientists, mathematicians and neurologists for AI research applications in healthcare. In the healthcare sector, for instance, AI is used in used in surgery by deploying robots, apart from its use in cancer treatment. Microsoft scientists are working on a project called Hanover in which AI will be used for genome mapping to strike at the roots of cancer. This is an elaborate research that aims to revolutionize healthcare by appreciating that cancer and other biological processes are information processing systems. The approach is about using tools that are used to model and reason about computatiol processes for modelling and reasoning about biological processes too. Now here you have an amazing mix of computer science and biology! How many research institutions in India are pondering on those lines?

The point here is not whether the Niti Aayog’s ambitious AI scheme for development is what the country needs at this time, given the other more pressing socio-economic issues the country faces. There is no use being a cynic too, just for the sake of opposing the budget proposal on AI for tiol development. The point is simple: the proposal is right and its time had come, but we first need to do our homework well by identifying the institutions that bear the potential for high-end, application-driven research. Then there is the great question of inter-discipliry approach, which is gravely lacking in our universities and research institutions. For instance, in banking and fincial services, how about a blend of economists, magement experts, statisticians and computer scientists sitting across the table and evolving AI research paradigms? This may serve as a serious food for thought for the Modi government. One only hopes something tangible would emate in the coming years.