Show caption Despite much hype about the threat of automation, it made little progress before the advent of Covid. Photograph: Alamy Automation can be our friend, but we must not let it turn into a foe Robert Skidelsky The growth of mechanisation brings many benefits, yet vigilance is needed to keep it in check Sat 25 Sep 2021 19.00 BST Share on Facebook
Share on Twitter
Share via Email
What the economic historian Aaron Benanav calls the “automation discourse” has been going ever since the luddites smashed textile machinery in Nottingham in 1811.
At issue is whether machines destroy or create jobs. The first case is easiest to understand. Machines are labour-saving; and labour saved means labour unemployed. Fear of unemployment has always been the dominant response of the workforce to the introduction of machinery.
The second case involves taking into account repercussions. The cheaper it is to produce something, the more demand there will be for it. This means more workers can be employed.
One can see then, how the spread of mechanisation to all branches of industry can multiply favourable effects: more people employed producing more and varied goods at higher wages for reduced effort. The fear of unemployment, say economists, is really a displaced fear of leisure.
With computer technology, not just physical work, but so-called “cognitive” work can be automated. Modern luddites foresee the growth of white-collar and service-sector unemployment. Again, say the optimists, they fail to notice the upside. The economic argument is straightforward: “Higher productivity implies faster economic growth, more consumer spending, increased labour demand, and thus greater job creation,” Sir Christopher Pissarides and Jacques Bughin argued in their 2018 paper.
The problem is social: to ensure that the fruits of increased productivity are passed on to the mass of the people in the form of higher wages and non-work income. The political debate is about how much public intervention is needed to ensure that the wealth created by machines trickles down to all sections of the population.
The interesting question right now is: what effect will the Covid-19 lockdown have on this automation discourse? Three effects in particular are worth noticing. The first is the likely speed-up in automation; the second, the increase in automatic shopping; and the third, the growth of home working.
Despite all the hype, automation made little progress in the UK before the pandemic. According to the International Federation of Robotics in 2018, the UK had only 71 robots to 10,000 workers. The main reason, I think, was that cheap labour from abroad was an alternative to automation, especially for small- and medium-sized enterprises that could not afford the capital cost of installing machinery.
However, this supply has dried up, and will not be rapidly restored. We now have the paradoxical combination of a near-record adult employment rate together with the highest ever job vacancy count.
Covid-19 is almost certain to accelerate automation in line with the experience of past pandemics such as Sars in 2003, with the driving forces being economic recession and the need to cut labour costs, and the perceived increased risk of human contact. Jobs with higher levels of physical proximity, such as retail, hospitality, leisure and medical care, are the most likely to be automated post-pandemic.
Unless the government intervenes to subsidise investment (say, through a national investment bank) the financing of automation will be brought about through a further concentration of industry in large firms and the bankruptcy of many small and medium-sized enterprises.
The second source of automation will come from the consumer switch to remote shopping. This is the joint result of a change in habits forced by lockdown and fear of contamination. One symptom is the rise in cashierless stores. The first Amazon Fresh convenience store (with automated sensors to detect when items are taken from shelves and which automatically charge customers) opened in London in March, promising a more “frictionless” consumer experience. Many more are promised.
Automation is only a means to an end. We need always to have in mind what purposes it is designed to serve
Finally, increased home working will demand increased use of surveillance technology. The proportion of working adults who did any work from home grew from 27% in 2019 to 37% last year on average, with Londoners the most likely to work remotely. Business sees a clear productivity gain in the reduction of time spent travelling and chatting in offices.
However, the realisation of such gains requires investment in surveillance technology. A recent report in the Financial Times highlighted the growth of electronic techniques for monitoring home working, including the installation of cameras and microphones in every house. This widens the discussion to the impact of technology not just on jobs but on freedom. When Jeremy Bentham invented his panopticon for monitoring the movement of prisoners, he suggested that it might be fruitfully applied in schools and hospitals. George Orwell carried this thought to its logical conclusion in his futurist novel Nineteen Eighty-Four. A two-way television screen in every flat ensured that “Big Brother is watching you” the whole time.
So on which side of the optimism-pessimism divide does the automation discourse now fall? Automation is not good in itself; it is only a means to an end. We need always to have in mind the question of what purposes it is designed to serve.
Unless this question is continually asked, and answered with action, we are destined to become slaves to the machines and those who control them. The luddites understood this.