When Andy Puzder, chief executive of restaurant chains Carl’s Jr and Hardee’s, said in March that rising employment costs could drive the spread of automation in the fast-food sector, he tapped into a growing anxiety in the US.
From touchscreen ordering systems to burger-flipping robots and self-driving trucks, automation is stalking an increasing number of professions in the country’s service sector, which employs the vast majority of the workforce.
Two-fifths of US employees are in occupations where at least half their time is spent doing activities that could be automated by adapting technology already available, according to research from the McKinsey Global Institute. These include the three biggest occupations in the country: retail salespeople, store cashiers and workers preparing and serving food, collectively totalling well over 10m people.
Yet evidence of human obsolescence is conspicuous by its absence in the US’s economic statistics. The country is in the midst of its longest private-sector hiring spree on record, adding 14.4m jobs over 73 straight months, and productivity grew only 1.4 per cent a year from 2007 to 2014, compared with 2.2 per cent from 1953 to 2007. Those three big occupations all grew 1-3 per cent from 2014 to 2015.
John Fernald, an expert on productivity at the Federal Reserve Bank of San Francisco, calls it the “Waiting for Godot phenomenon”, referring to the Samuel Beckett play. The US may be home to the most prodigious centres of technological innovation on earth, but the results in terms of efficiency gains remain elusive.
“You would think if you were seeing robots taking over, we would see lots of capital growth and not much labour growth,” he says. “At this point it is not pervasive enough to be having a big effect.”
That is not to dismiss the impact new technologies can have on professions. Western nations have been automating for hundreds of years and the process has led to painful disruptions, prompting economist John Maynard Keynes to write in the 1930s about the threat of “technological unemployment”.
Agriculture, for example, employed nearly 60 per cent of the US labour force in the mid-19th century; by 2000, the sector employed about 2 per cent. In US manufacturing, the number of jobs has fallen by almost 5m from the start of this century, in part because of automation, as well as trends including the economic integration of lower-cost markets such as China.
“The fact that the percentage of men aged 25-54 employed in the United States slowly but steadily declined since the 1950s, as manufacturing has shifted to services, suggests that challenges may arise,” the Council of Economic Advisers noted in its annual report this year.
Mr Puzder says sharp increases in the minimum wage in states such as California and New York, coupled with other employee costs including healthcare, will prompt greater use of automation in restaurants. “The fact of the matter is automation is coming,” he says. “The question is how fast is it coming.”
McKinsey’s data show that nearly three-quarters of work done by workers preparing and serving food in restaurants, numbering more than 3m people, could be done by robots.
That does not mean three-quarters of those workers could suddenly find themselves replaced by computers, however. Just because it is technologically feasible to replace a set of human tasks with work done by a robot, that does not mean it is cost-effective, especially in lower-wage sectors, says David Autor, an economist at MIT.
Personal services, in particular, are difficult to deliver via robots because they require “flexibility, adaptability and common sense”, and they are expensive to automate.
Technology companies themselves are unsurprisingly eager to scotch the notion that their products will destroy the US jobs market.
Sarah Doherty is chief technology officer of TeleHealthRobotics, a start-up based in Chicago that is testing a robot that performs ultrasounds on patients under the supervision of a doctor at a remote site. From helping lift heavy patients to moving supplies through corridors and assisting in consultations, robots will play a growing role in healthcare, she says.
But she tries to reassure healthcare professionals that companies such as hers are not trying to supplant them. “Our feeling is it can be an aid and a supplement and not a threat,” she says.
Michael Chui, a McKinsey Global Institute partner, says it is important to examine which jobs are most susceptible to automation, given it may simply be taking time for transformative technologies to be adapted and rolled out across workplaces. While new roles may emerge, the question is whether they appear quickly enough to replace the lost jobs. “It is extraordinarily difficult to predict where new job creation will come from,” he says.
The hollowing-out of middle-skilled white collar and manufacturing jobs because of computer technology has already pushed many Americans into poorer-paying professions, contributing to the wage stagnation and inequality that are convulsing the political scene.
Andrew McAfee, an MIT scientist and co-author of The Second Machine Age with Erik Brynjolfsson, says automation is starting to be applied to less-controlled environments, such as distribution warehouses, expanding robots’ footprint beyond the realm of repetitive tasks. Jobs that involve pattern matching, including customer service and radiology, could also be affected.
More than 40 years after the introduction of the first ATM, the number of bank tellers is finally declining in the US, Mr McAfee points out. “Examples [such as these] convince me that after we have enough kinds of automation that can do portions of someone’s job we can see the point where we pass peak total demand for that job,” he says.
Nonetheless, “it is easy to underestimate the ability of innovators and entrepreneurs to figure out new things that we want and hire people to deliver them.”