The robots are coming – but not fast enough

The claim that mechanisation is sweeping away jobs in a wave of innovation bears little relation to reality.

James Woudhuysen


In August, Japan hopes to send Kirobo, a 34-cm tall, 1kg humanoid robot, to the International Space Station, to assist a Japanese astronaut. Robots are getting cleverer, more dextrous and safer; interest has turned to them performing services, working alongside human beings, and maybe eventually acting in ‘swarms’. But aren’t robots also destroying jobs?

Management consultants McKinsey say they are. McKinsey argues that, by 2025, robots may take 15 to 25 per cent of developed-world tasks in occupations such as manufacturing, packing, construction, maintenance and agriculture, and five to 15 per cent of developing-country tasks in manufacturing alone. It also forecasts a diffusion of robots in commercial services at about half that rate.

McKinsey expresses a growing consensus. In America, top Keynesian economist Paul Krugman declares ‘sympathy for the Luddites’, arguing that robots now hit educated workers of all descriptions. Krugman also believes that robots sway the distribution of national income further towards capital and away from labour. Others make similar points about automation in general.

Everywhere, the advancing capabilities and spread of IT are felt to increase inequality. McKinsey warns of ‘social and political resistance, particularly if robots are perceived as destroying more jobs than they create’, and favours extra education for those made redundant by robots. Krugman wants a state-welfare safety net for those whose jobs are robotised.

Is any of this a realistic account of events? For the penetration of world robots to meet its forecasts for 2025, McKinsey concedes that a lot of investment will be required: about $1 trillion to increase industrial robots from 10million to 25million, and a further $0.2-4 trillion to up commercial service robots to 2.5-2.8 million. That, McKinsey adds, would imply world robot sales growing by 25 to 30 per cent every year over the next 13 years – ‘significantly higher than the average growth rate over the past two decades, but lower than the growth rate in 2010 and 2011’.

Er, but what about the recession? The word doesn’t appear in McKinsey’s report. That’s odd. In the West, company cash hoards have swollen, while many research-and-development (R&D) budgets contract. Female, youthful and immigrant labour are cheap. So why, during an investment slump in the US and elsewhere, would employers rush to spend money on machines that demand radically new processes and expertise? Robots can cost $15,000 if you buy 1,000 of them, but organising a manufacturing cell around each one will add 10 times that to your bill.

Elite commentators would rather believe that robots accentuate income inequalities than check what’s actually happening. In 2012, sales of industrial robots in North America hit a record 22,598 machines. But suppose, generously, that each robot replaced 10 jobs: then the number of workers displaced over the whole of 2012 would nearly be compensated for by the number of new jobs sluggish America and more dynamic Canada currently create each month.

For all this, though, debates about whether robots can ever create as many jobs as they destroy are misplaced. By itself, even the most powerful innovation cannot make or nullify a single job: it is human beings, and the rhythms of capital accumulation, that do that. And right now both the temper of human beings and the tempo of capitalism do not favour robots.

Those unnerved by robots, after all, don’t just exaggerate their impact on employment. They also drum up the dangers mobile robots represent to safety. They fret about the ethics of using robots – especially in the home, and especially with older people. And they write about robot ‘smog’, describing it as a ‘spectacular zoo’ which might ‘imprison us’.

It’s true, in key sectors such as electricals/electronics and automotive, that the number of robots per industrial worker grew, 2008-11. However the pattern is spotty:

Internationally, there’s little or no consistent, durable or ‘exponential’ trend towards the rapid adoption of robots. Taiwan’s Foxconn Technology Group has 20,000 robots and plans nearly a million more; in Korea, surveillance and security robots made by Samsung Techwin stalk the North/South demilitarised zone. However, both in 2011 and 2012, just 160,000 industrial robots were sold worldwide, while average annual increases in global robot sales were nine per cent, 2002-12. That is an impressive rate of growth; yet it is just a third of what McKinsey imagines for 2013-25.

Robots have long gripped the human imagination. Today’s unprecedented anxiety about them doesn’t much reflect recent advances: though tangible, improvements in robot hardware and software remain limited. Nor does robophobia reflect a substantially faster diffusion of robots, or any genuine epidemic of mass, robot-induced unemployment.

In effect, robophobes fear that catastrophe is upon us, even though not much is happening. Worse, they credit capitalism with much more innovation than, right now, it can actually muster.

James Woudhuysen is professor of forecasting and innovation at De Montfort University, Leicester, and blogs on

To enquire about republishing spiked’s content, a right to reply or to request a correction, please contact the managing editor, Viv Regan.