Hide the batteries, unplug the router, learn to code: if there’s a truth universally acknowledged, it’s that robots are coming for your job. According to this familiar story, our economic future will be one of mass unemployment. Automatons will inevitably take the place of human workers. These robotic employees won’t complain, suffer fatigue, or unionise.
The new robots will be “white collar” predicts Richard Baldwin, author of the recently published Globotics Upheaval: Globalization, Robotics, and the Future of Work. (“Globotics” is a clunky portmanteau of “globalisation” and “robotics”.) Western professionals, no matter how qualified, will be supplanted by low-wage foreigners working remotely via Skype and UpWork, and highly-tuned algorithms that can out-think human beings. Baldwin’s advice? Don’t compete with robots, or “tele-migrants” for that matter. Find the competitive edge in your own humanity. Become a Zumba instructor.
Yet this narrative is dubious. An influential 2013 Oxford University paper suggesting that 47 per cent of US jobs were vulnerable to automation remains widely cited, but when its methods were replicated with different weightings, researchers obtained radically different results. A recent paper by the Institute for Public Policy Research (IPPR) found that the number of US jobs that could be wholly automated was relatively low: an estimated 5 per cent. Two other recent studies – one by the Organisation for Economic Development (OECD) and one by the Roosevelt Institute – both suggest the consequences of automation depend more on politics than robotics. Far from being inevitable, the outcomes of new technology will be determined by conscious policy choices.
Automation anxiety has a long history. In the 1930s, John Maynard Keynes warned of a new disease – “technological unemployment”. In 1961 US president John F. Kennedy established the Office of Automation and Manpower to study the “major domestic challenge of the 60s: to maintain full employment at a time when automation, of course, is replacing men”. As fears of technological obsolescence grew, his successor Lyndon B. Johnson inaugurated the National Commission on Technology, Automation, and Economic Progress to examine the risk of robots usurping human labour.
Yet automation isn’t a linear process. The vacuum cleaner and washing machine have shown how new inventions destroy some jobs and create others, in part by reducing the burden of household labour. The IPPR paper found that automation generates more jobs in the long term than it eliminates – while also correlating with increasing productivity, as it takes fewer hours work to produce the same output. It’s true that contemporary digital capitalism differs from previous incarnations, and that technological progress has accelerated (it took seven decades for phones to penetrate households, but only nine years for smartphones to catch on).Yet technologies don’t exist in a void. Their introduction frequently exacerbates pre-existing underlying power dynamics within the economy.
Authors of the three papers agree on one point: new technologies could increase prosperity, but whether or not they do so will depend on how their economic fruits are distributed. If the robots ever come, what they are used for and who profits from them will be determined by economic distribution and property ownership regimes.
So why does optimism about the astonishing progress and transformative nature of robots persist? In part, because headlines blaring job losses drive more clicks. As the authors of the Roosevelt Institute report note, however, the reality is more complex. Were a robot revolution truly underway, we might expect to see an increase in worker productivity. But large-scale productivity gains continue to elude Western economies. Robot enthusiasts attempt to explain away this paradox: either economists are mis-measuring output, or new technologies are overly confined to selected firms, or productivity increases haven’t shown up in the statistics – yet.
The reality, though, might be that new digital technologies aren’t disrupting the economy as much as their advocates suggest. Workers are right to be worried about new technologies – but not because they’re about to be replaced. More than the spectre of job obsolescence, it’s the consolidation of corporate power that haunts the labour market.
The history of management is partly one of attempting to measure and control work. Technologies that mimic human capabilities may further empower employers in a labour market already tilted in their favour. Trade union researcher Victor Figueroa observes that, just as unscrupulous employers have used microphones and email monitoring to pre-empt union organisers, Artificial Intelligence – which is already used to sift through employees in hiring processes – could also be deployed to create personality profiles to pre-empt strikes and identify insubordinate staff. In warehouses around the world, many workers already complain of inhumane conditions when algorithms set the pace of work.
AI and machine learning threaten to exacerbate existing trends, creating a new digital overseer at a time when union membership stands at a near-record low: union membership in the UK has plummeted from 13.2 million in 1979 to 6.2 million in 2017; in the US a mere 10.5 per cent of workers are unionised. As the OECD report observes, while anxieties over automation abound, the greater risk is of widening disparities between workers. Some will have every aspect of their labour measured and analysed, while others will have the power to avert this fate, Figueroa predicts. Dystopian visions of advancing robots shouldn’t distract from the more familiar struggles we face.