The automation of work through artificial intelligence (AI), robotics, 3-D printing and other new technologies has been a hot topic in recent years. But it has now got a second wind, courtesy of the Covid-19 pandemic.
The Future of Jobs 2020 report published last week by the World Economic Forum (WEF), famous for its annual jamborees of the world’s movers and shakers in the Swiss resort of Davos, points out that automation, combined with the pandemic, is creating a “double disruption” scenario for workers.
Besides inducing a global recession that is fuelling unemployment, Covid19 has also accelerated companies’ plans to automate work processes, which could lead to even more job losses. In its October regional economic outlook for the Asia-Pacific, which was released last week, the International Monetary Fund (IMF) echoes this broad conclusion, noting that the deployment of robots and other labour-saving technologies tends to increase both after pandemics and during recessions – and we are in both.
Asian economies are especially vulnerable, it points out, given that about two-thirds of the world’s operational stocks of industrial robots are in Asia, and that robot density is rising fast, from a low base in several Asian economies, many of which have a high dependence on labour-intensive manufacturing.
Think of the garment industry in Cambodia and Bangladesh, or the consumer electronic assembly plants in Vietnam and the Philippines, or the car factories in Thailand and India, not to mention the vast agricultural sectors across the region, which are also vulnerable to automation.
The news is not all bad, at least in the medium term, according to the WEF, which estimates that while 85 million jobs worldwide will be displaced by 2025, 97 million new roles will emerge. This is consistent with the usual outcome of technological advances, which have historically ended up creating more jobs than they destroyed.
But the story may be more complicated this time. There are many uncertainties and bumps along the way between now and 2025.
For one thing, amid the pandemic, job creation is slowing while job destruction is accelerating. Automation (which will add to job destruction in the short term) is easier than in the past. For instance, AI algorithms can be easily shared and deployed. Robots can be quickly installed in existing factories. So can 3-D printers; no major new infrastructure is required.
The rise of remote work, the demand for which is growing dramatically, has created further complications. The job losses are concentrated in relatively “high-contact” and “non-teleworkable” industries such as hospitality, retail, mining, manufacturing and construction. The most affected are lowwage, less educated workers and youth as well as women, many of whom have been forced to stay at home to manage child-rearing and household responsibilities, especially during lockdowns. Skill gaps – the difference between the skills that are available and those in demand – are high.
The WEF report points out that there is a growing demand for skills that include critical thinking and analysis, problem-solving, self-management, resilience, stress tolerance, creativity, originality, the ability to work with people as well as technology, management and communication skills and flexibility.
The new job openings are in areas that include AI, data science, cloud computing and the green economy, but also various sales jobs, management and administrative roles, and jobs in social sectors such as health.
In Singapore, the three top roles for which employers need workers, according to the WEF report, are data analysts and scientists, AI and machine learning experts, and digital transformation specialists.
Jobs that are becoming increasingly redundant as a result of automation include data entry clerks, accounting, bookkeeping and payroll clerks, administrative and executive secretaries, and accountants and auditors. The WEF is not altogether pessimistic about the capacity of displaced workers to make the transition into at least some of the jobs in high demand.
Data from LinkedIn – which the report drew upon – indicates that datarelated, AI, sales and even some engineering jobs have not proved difficult to break into.
Data from the report also shows that during the pandemic, workers are already making job and career transitions into unfamiliar areas. For example, around the world, many workers displaced from hospitality, retail, food and beverage (F&B) and entertainment industries have joined the health sector; retail and hospitality workers have become bankers, some from higher education have become software engineers, and F&B workers have become teachers.
Nevertheless, the WEF report points out that given the accelerated pace of both job losses and automation, “the window of opportunity to reskill and upskill workers has become shorter”.
Therefore, there is an urgency for governments to monitor the new opportunities emerging in labour markets and support the retraining of displaced and “at-risk” workers, including through vocational programmes and online training platforms. The IMF also calls for governments to revamp educational curricula and encourage lifelong learning with a view to building more flexible skill sets. Many observers note that Singapore has been one of the early movers in all of these areas, having launched training subsidies and lifelong learning initiatives long before Covid-19.
But over and above this, both the WEF and the IMF call for an expansion of social safety nets to protect the unemployed and to reduce inequalities there is a high risk that job losses will be far more rapid than job creation, even with the best efforts put into reskilling, and inequalities are likely to rise. This is already evident in Singapore, where data from the second-quarter Labour Market Report shows that blue-collar workers have borne the brunt of the layoffs and furloughs.
The WEF report has its limitations. It is mainly based on feedback from a few hundred large companies – the WEF’s corporate constituency. It did not survey small and medium-sized enterprises or firms in the informal sector, which employ far more people globally (and also in Singapore) and are both more likely to lay off workers and less likely to accelerate automation.
So, its projections of the total numbers of job displacements and new opportunities are based on a limited sample with a big-company bias.
Some economists are doubtful that the new opportunities will necessarily exceed displacements. Although this has been historically largely true of automation, it has not always been the case; history is not a reliable guide. For example, in a recent study, economists Daron Acemoglu of the Massachusetts Institute of Technology and Pascual Restrepo of Boston University found that within industries adopting automation in the US, the average “displacement” (or job loss) from 1947 to 1987 was 17 per cent of jobs, while average “new opportunities” created was 19 per cent.
But from 1987 to 2016, it was a different story. The displacement was 16 per cent of jobs, while new opportunities represented 10 per cent. Not only that, but beneficiaries of the two separate waves of automation were different. Professor Acemoglu pointed out that technological changes from the 1960s to the 1980s benefited mainly low-skilled workers, but the automation thereafter, and especially since the 1990s when robotics began to take off, benefited more highskilled workers, while displacing many of those with low skills.
Higher-skilled workers are also likely to be bigger beneficiaries of the new wave of AI and software-driven automation. Prof Acemoglu cautions that while many technologies are job displacing, they are not necessarily job-creating. He gives the example of self-checkout counters at supermarkets and self-check-in kiosks at airports, which he calls “so-so technologies”. They reduce costs for firms a little bit and displace some workers (because the customers do the work) but they don’t increase productivity by much, or greatly benefit other workers, or lead to much job creation.
It is important therefore that companies are not encouraged to adopt technologies for technology’s sake, but to adopt those that are job-and productivity enhancing rather than just job replacing.
This is not always easy to predict because later innovations can disrupt earlier innovations. For example, the Internet and file transfers enabled radiologists to interpret X-rays remotely, yielding cost savings for hospitals and clinics and new jobs for radiologists in remote locations. But the later arrival of AI – which can interpret X-rays more accurately than humans – disintermediated the remote radiologists.
Similarly, Global Positioning System technologies enabled ride-sharing, creating jobs for millions of freelance drivers. But autonomous vehicles (in which ride-sharing companies like Uber are investing) will displace those drivers.
Even data scientists and data visualisers, who have been among beneficiaries of the big data revolution, could be displaced by machine learning. So rapid is the pace of technological advance that which jobs are potentially at risk can be a tricky conundrum for individuals, companies and policymakers.
Just as important as keeping up with the curve of innovation will be to train for flexible skill sets, and to have strong safety nets to protect displaced workers, because there may be many of them, in several waves to come.
The Straits Times