Some limits to automation

Predictions about the future often overestimate how far current trends will hold.

I read Martin Ford’s sensationally-titled The Rise of the Robots last week, and couldn’t help but feel that it falls into this trap. It wasn’t Ford’s analysis of technological trends that I took issue with: his prophecy of imminent mass unemployment is reasonable on the basis of the capabilities of today’s technology, let alone the machines of the future. But just because a robot can do a job doesn’t mean that it always will. My sense is that a backlash on the part of workers and consumers (and with them, politicians) could act as a hard brake on the mass automation of labour before long.

The most inevitable area of resistance will be on the part of workers themselves. Industrial action contesting large-scale redundancies may become commonplace, with the interminable strikes hitting Southern Rail in recent months a portent of what’s to come. Many strikes will prove unsuccessful, and those involved might go on to acknowledge the benefits brought by the new technology further down the line. But some combination of disruption to business and bad PR — with public sympathy for strikers ballooning as increasing numbers of people come under similar pressures — could see companies rein in their automative tendencies after a certain point, rather than accelerating them.

Unemployed workers will also vent their frustration through the ballot box. The popularity of Trump’s promise to return manufacturing jobs to the US could herald a new norm, with the much-reported rise of populism across Europe part of a similar trend. We’ve seen what people think about ‘foreigners’ taking their jobs, but how different will it be if those foreigners are kicked out and machines take the jobs instead? Elections may be fought and won on the basis of rolling back the tide of automation in the very near future.

A political backlash against the computerisation of the workforce will be further bolstered by practical macroeconomic considerations. Too many humans falling out of work means declining tax revenues, a heavier burden on the welfare state, and an economy incapacitated by a plunge in the purchasing power of the people. Push-back against automation is likely to be significantly more politically feasible as a solution to these problems than alternatives, such as the introduction of a universal basic income.

The question of purchasing power also highlights a limitation, government interference aside, on companies who might be looking to automate — and one that Ford does acknowledge, to his credit. It’s all very well cutting jobs to boost profits, but how do you make any money if those job losses leave the mass market decimated? There must come a tipping point after which automation stops working in most companies’ favour. Combined with the scope for industrial action and changing public opinion noted above, that tipping point could arrive sooner than we think.

It’s in workers’ capacity as consumers, too, that we could see a backlash against automation. This might be on any number of fronts. If consumers feel threatened by robots in their own workplaces, they may favour businesses that employ humans over machines — doing so could even come to be seen as a form of corporate social responsibility. A recent Nesta article argued that a market for work explicitly not done by computers might burgeon in the coming years, as people become disquieted by the ubiquity of algorithms, and distrustful of the decisions they make.

And what of roles that currently involve face-to-face interaction? To Ford’s mind, shop assistants, for instance, will be some of the first to lose out. But in Japan, the country often picked out as being most open to robots in daily life, smiling shop assistants abound, and even self-service checkouts are largely absent. The Japanese national Robot Strategy actually emphasises the human role in hospitality, with robots in the service sector of the future consigned largely to storerooms. Perhaps this is cultural. But it may also be that as computers become ever more prevalent in our lives, we come to appreciate human interaction just that little bit more. In that case, it might be investment in the training of humans, rather than algorithms, that continues to decide a company’s fate.

I’m not trying to say that jobs won’t be lost in the coming years: many will, and we do need to think about how to negotiate the issues that will result. But, even if technologically possible, there is nothing inevitable about a future in which up to half of the workforce has been replaced by computers. A backlash against automation might yet remind us that technology is not the only driver of social change.

For a few other thoughts on the topic + a brief reading list, see here.