News about the miraculous things that Artificial Intelligence (AI) is about to unleash is greeted with scepticism even by technophiles because, it is one of those things that since the 1950s has periodically risen making large claims about its earth-shattering promise only to bashfully slip back into oblivion. Why then are, normally sensible news outlets such as the The Telegraph of London, quotes another expert as saying that "In five years most companies will have armies of very cheap, very powerful machines, connected to a smart cloud, and they will be able to accomplish so many routine processes… that is presently being done by human beings thereby destroying millions of jobs".
What exactly is AI about which all this hullaballoo is being created about? You may already be a user of some of the simpler forms of AI: When you finish reading a news article on a news site and click on a link that says "more like this" to get to another article on the same topic you have already taken a tiny step as a user of AI. Behind the scenes on that news site, a computer programmer has written a code that has analysed all news articles on that site and grouped them as belonging to "topics" such as "Bollywood Film Reviews" or "Indo-Pak Relations". So a reader who has evinced interest in one Bollywood Film Review is presented links of other Bollywood Film Reviews.
Driving cars, which involves co-ordination of your hands, feet, eyes and ears as you weave in and out of traffic is currently a favourite playing field for AI, particularly in countries such the US, where a very large proportion of the population are more than 70 years old, lives alone and has to drive itself. The degree of automation could be as basic as automatic braking or could extend to putting a car on autopilot, and could soon extend to truly driver-less cars. All this is accomplished by devices in the car that detect its surroundings using age-old technologies such as radar and GPS (that's what tells you your location on your smartphone), to detect obstacles and road signs, and software that uses this information to identify navigation routes.
Such advances are not without its risks. A driver-less car that can find its way through complex traffic can also be a deadly tool for terrorists who could either directly use these vehicles to carry bombs to deadly effect. Warfare could get radically changed with both parties to a conflict fielding their respective driver-less tanks. Early examples of this kind of warfare are the use of drone (pilotless) aircraft to make strikes in remote areas. Then, there is the risk that many such advances as described above could result in the displacement of human beings who are currently performing such tasks.
A panel of top academics and industry practitioners has been formed in the US to assess where AI technologies are going and their impact on society by 2030 and has just issued its report (Shivaram Kalyanakrishnan of IIT-Bombay is a member of that panel).
In India, it may be important to ask what ought to be the response of India's public policy establishment to these technologies? As a first step it may be fruitful to shortlist some areas in delivery of public services that can benefit the most from AI and make sure that product development and deployment efforts are easily financed for these efforts. Some areas could be education (tools to make the task of teachers in primary schools, industrial training institutes and polytechnics easier), healthcare (tools to enable the primary health care system more easily diagnose and prescribe remedies) and national security (early warning systems for terrorist actions), to name just three.
The writer is the author of The Wave Rider, a Chronicle of the Information Age.