WHEN a Google computer programme called AlphaGo beat Fan Hui, the European champion of Go last January, a wave of excitement ran through the geek community. Actually, there should have been a ripple of concern among social planners and economists as well.

Go is a fiendishly complicated game invented in China 2,500 years ago, and is played by 40 million players around the world. It has 200 possible moves compared with 20 for chess, and has more possible board arrangements than there are atoms in the universe. So it is hardly surprising that most experts were convinced that it would take at least a decade more for a computer to beat a good human player. In fact, AlphaGo humiliated Fan Hui 5-0.

Impressive though the feat is, how does it concern those of us who don’t play Go, or interest ourselves in advances in computer science? It might not be obvious, but there are reasons for future generations to be afraid, very afraid.

According to forecasts, around a third of all jobs in the UK will be lost to robots and computers by 2035.

The brain behind AlphaGo’s victory was Artificial Intelligence (AI), the Holy Grail of computer scientists. Most computers today, although far more powerful than their predecessors, are still very stupid, and will only do exactly what they are programmed to. No more and no less. So they can’t cope with unexpected moves by an opponent. AI, however, is capable — to an increasing degree — of mimicking the way a human brain functions by creating neural networks. Most importantly, AI-enabled computers can learn from their mistakes.

When IBM’s Deep Blue supercomputer beat Gary Kasparov, the chess world champion, in 1996, it did so largely by using brute computing power. Typically, a chess-playing computer has a large number of games stored in its memory, and responds to a move by analysing all possible moves and ranking them in order to find the best. At the start of a game, this is relatively simple, but as the game develops and opens up, the number of possible moves increases exponentially. And the computer has to project future moves, according to its opponent’s responses. Millions of calculations are thus necessary for every move.

A human player, on the other hand, will simply reject the obvious rubbish moves, and focus only on the most promising line of attack or defence. This inherent superiority, we thought, would give humans a permanent edge over computers. Scientists are no longer so sure, given the advances made in AI. Already, cars pioneered by Google are cruising the streets of California, guided entirely by computers. Other car-manufacturers are hot on its heels, and some think that by 2020, many American states and European countries could legalise driver-less cars.

The implications are enormous. Roads would be far safer as human error accounts for the vast majority of all accidents. But equally importantly, most of those who drive for a living will be out of jobs. A tap on an app would summon a self-driven car. Indeed, car ownership would gradually become a thing of the past. Taxi and truck drivers, as well as private chauffeurs, would become unemployed. This is not science fiction, but our immediate future.

The Japanese were early pioneers in making industrial robots, and millions of these machines are now on factory floors across the world. But with their ageing population, Japanese scientists have also made robots to help with housework and care for the elderly. One machine can be programmed to vacuum rooms; another mows lawns. Both are equipped with sensors that prevent collisions with randomly placed objects.

Middle management jobs and those requiring repetitive operations are first in line for the chop. More will follow. According to forecasts, ultimately the top jobs and the nasty, dangerous work will be left for humans. Millions will thus be out of work. Already, unemployment among the young in advanced countries is high. How will the new world look like in a couple of decades?

Clearly, such a future will cause major headaches for governments and planners. And yet, a world of robots is inherent in the capitalist model: machines lower operating costs, maintain uniform quality, and above all, there are no troublesome unions to deal with.

But workers are not just figures on a production line: they are also consumers who spend their wages on goods manufactured in other factories. If they lose their jobs, who will buy all the items offered in shops and showrooms? Clearly, the implications of a world of AIs have not been thought out.

In the Third World, however, labour is still so cheap that robots are a long way away. Nevertheless, those manufacturers wishing to compete in the global marketplace will have to adopt the latest technology. This will mean some unemployment among skilled workers.

In the West, growing numbers are now classified as self-employed, and lifelong employment in one firm is now a thing of the past. Generally speaking, each industrial revolution in the past has created new jobs. Thus, when steam engines, and then electricity, created a new industrial landscape, new categories of work emerged to absorb the workforce that had been made redundant. Similarly, when computers eliminated vast numbers of jobs, new employment possibilities opened up.

But this time, social scientists think there will be no similar job-creation following the AI-led industrial revolution. So while the future looks exciting, it is also a cause for concern. I’m just sorry I won’t be around when cars drive themselves, and robot chefs turn out wonderful dishes from thousands of recipes stored in their memory.

Twitter: @irfan_husain

Published in Dawn, February 8th, 2016