Last week, 20 years after a computer named Deep Blue defeated the then world chess champion Garry Kasparov, a machine called AlphaGo beat Lee Sedol, the best player of Go, a game more complex than chess and popular mainly in East Asia. While that massive blow was being struck for artificial intelligence, I was attempting to get humans employed by the most respected private sector bank in India to answer a simple question about my credit card. They seemed to be trying, but were restricted either by company policy or personal limitations. I got emails from Uvaraj, Padma, Melkis, Sathish, and Gayathri, but without exception they provided cookie-cutter responses culled from some list of FAQs rather than tackling the specific issue I had raised.
If a machine could beat the best human players of Go, I thought, surely one could be designed to answer my queries better than my bank’s customer-service department. As it happens, such a machine already exists, and its name is Watson. Created by IBM as the next challenge to their artificial intelligence initiative following their world-conquering chess programme, Watson is a computer system designed to answer questions posed in colloquial language. It made a splash five years ago by beating previous winners of the popular quiz show Jeopardy, and has already found commercial application in sectors like healthcare, financial services, legal research and telecommunications. As Watson gets more powerful and less expensive, I can see it replacing the likes of Uvaraj, Padma, Melkis, Sathish and Gayathri.
Robotic process automation is already a buzzword in industries around the world. It presents a massive challenge to India’s software services sector, alongside the shift to cloud based services and away from individual data centres maintained by firms’ IT department augmented by outsourced manpower. Indian software firms understand very well a revolution is underway. Wipro has spoken of trimming its workforce by 30%. Infosys hired as its Managing Director Vishal Sikka, who has a Ph.D in artificial intelligence from Stanford. The perspective of privately owned corporations, though, is exclusively geared towards increasing profits. Shifting focus from a staff augmentation model to a less labour intensive one is a technical challenge for them, but not a moral one. That’s the way it should be, for no company can retain staff as it slides into bankruptcy.
Social disruption
Framers of public policy, however, must look at the issue from the perspective of human employment. A drop in the number of professionals working in the software services sector would be severely disruptive to the nation’s social fabric, and I don’t use the word disruptive in the positive sense it has gained in the tech world. If call centres becomes obsolete, if the BPO business is decimated, and if a range of rote jobs across the service sector fall prey to automation, as is more than likely to happen over the next decade, it will cut off what has been for the past two decades the most important pathway to affluence for educated middle-class Indians.
It was inevitable that automation would at some point push India’s software exports sector towards greater productivity. The industry has grown from its infancy exclusively by adding workers to increase earnings. The revenue firms make for each person employed has remained virtually static for years, and is abysmally low, at well below $50,000 per year per employee.
Indians were among the first people to feel the deleterious effects of technological progress, when the textile sector was mechanised two centuries ago. We were spared the disruption caused by the mechanisation of agriculture, depicted most movingly in John Steinbeck’s novel about the dust bowl, The Grapes of Wrath. We avoided the worst effects of improved efficiency in industries such as steel, efficiency which created the rust belt in the United States. We avoided it mainly because Indian companies are forced to keep on their payroll staff they no longer need. JSW Steel manufactures more or less the same amount of metal as the Steel Authority of India, but has a workforce only one-eighth the size of SAIL’s and a labour cost per tonne of steel one-sixth that of the public sector behemoth.
Blundering on
The technological revolution in telecommunications and the Web of the first decade of the twentieth century helped India even as it hurt workers in affluent nations by opening up the potential of offshoring. The current revolution is likely to be less benevolent. Many of the jobs that were Bangalored in the past decade-and-a-half will now be Watsoned. And no government regulation can prevent the haemorrhage.
Oblivious to these monumental changes, engineering schools have proliferated around the country, to the point that we now produce hundreds of thousands of mostly mediocre graduates each year. Even though the central regulator clamped down recently, we have a massive oversupply of engineers possessing only rudimentary skills. With software service providers needing to climb the value chain, there are going to be fewer opportunities for everybody but the very best grads.
India’s ITES industry employs between 3 million and 4 million people directly, which might seem like a drop in the bucket in a nation of over a billion. But consider that some 94% of India’s workforce is employed in the unorganised sector and that the government accounts for the majority of jobs in the organised sector. ITES provides one in every four jobs within the privately owned, organised sphere, which is a massive chunk, commensurate with its disproportionate contribution to the nation’s exports and GDP. Should employment in this sector fall, it will be like being shaken awake from a wonderful dream.
Recognising the danger
I’m not sure the government is adequately cognisant of the threat. Among its signature initiatives is "Skill India", which appears laudable in theory but is confused in practice. KPMG was tasked with projecting the future of employment opportunities in India. Its report on the software sector forecast an increase of 2 million jobs over the next seven years. This is an inordinately sanguine outlook, not shared by any expert I’ve read on the subject. When every major firm is speaking about cutting flab and projecting workforce reductions up to 30%, when many smaller companies seem unlikely to cope with the tremendous technical churn underway, and when corporations across the US and Europe have mandated cuts to their IT budgets, where are those two million extra jobs going to come from? It’s like somebody doing a survey of the steel industry in 1975 and predicting a 60% increase in the workforce over the coming decade.
As an aside, a second technological revolution threatens our best paid blue-collar workers. I am speaking of the six million Indians employed in countries in West Asia wounded by plummeting oil prices. So far, Gulf nations have dipped into their cash reserves to keep investment up, but austerity is around the corner, and around the next corner is recession, should crude prices stay low for years as they probably will. I believe that 30 years from now places like Abu Dhabi will be what Detroit is today, formerly glittering centres of wealth reduced to a shadow of their former selves. Migrant workers are going to get fracked in the process even as Bangalore gets Watsoned. What happens when much of the $35 billion remitted annually by Gulf workers dries up? More generally, what if the demographic dividend of which we are so happy reverts to the nightmare of overpopulation it was in the 1970s, with millions of young men and women growing frustrated and angry at the dearth of decently paid or reasonably secure employment?
I am not suggesting that doom and gloom are assured. New technologies could open up vistas nobody can see right now, creating new spaces for graduates as well as blue-collar labourers. But we should always factor in worst-case scenarios, and I don’t see anybody in the policy establishment doing that at the moment.