In the early days of automation, machines ruthlessly decimated workers’ place in the commercial order, and it did create jobs – with a twist. Many of those who were displaced were either too geographically immobile (they could not physically access these opportunities), or more commonly, too occupationally immobile (they did not have the necessary skills). Fortunately, increased dynamism in the workforce has dismantled these impediments, refashioning robotisation into a desirable phenomenon which generates higher wages for more people.
Counterintuitively, automation leads to a net gain in jobs; the World Economic Forum predicts that 2020 to 2025 will herald the creation of twelve million vacancies (Kande & Sonmez, 2020). Contrary to the suppositions of academics such as Kai-Fu Lee, a detrimental cycle in which robots take up these newly created roles every few years will not be instigated, for the foreseeable future. There will always be occupations which require the interpersonal interaction that non-sentient robots cannot replicate, such as caregivers and teachers. Furthermore, machines are vastly less dexterous than humans (they have less than forty times the number of axes of movement), hence will not be able to replace other groups, including mechanics and electricians. And as staggering as the rate of robot development may seem, automation will not replicate the hundreds of thousands of motor neurons in the human body any time soon.
More jobs translate into higher earnings because many of the 220 million individuals who were previously unable to find work may now do so, and workers’ bargaining power in salary negotiations grows as they can more readily threaten to quit for a higher-paying company, thus instigating a positive feedback loop. The reasons for robotisation leading to more jobs are twofold. The fact that operating a robot will cost 16% less than human sweat by 2025 (a disparity which is set to grow as machine capabilities improve, economies of scale develop, and the working class obtain more political capital to lobby for better working conditions) will render businesses more profitable (Young, 2015). Firms can thus expand and create more occupations, a process which will be accelerated by diminishing costs leading to lower prices, further demand for their products, and more revenue. For example, between 1988 and 2004, the introduction of ATMs decreased the number of bank tellers per branch by 35% but increased the number of branches by 43% and led to greater employment of salespeople (The Economist, 2016). In addition, automation requires hiring more workers, albeit for higher-skilled occupations, to look after these robots. This chiefly includes data analysts and machine learning specialists, but it will supercharge demand for corporate lawyers to boot as many regulatory questions will be raised, especially concerning the omnipresent debate on the privacy of data employed in robotic systems.
Furthermore, robots will generate exuberant amounts of economic growth by boosting productivity; not only do estimates postulate that they will contribute $15 trillion over the next decade (PwC, 2020), but also that between 1993 to 2007, they accounted for one sixth of productivity growth, on par with the successes created by railroads and US highways (Michaels & Graetz, 2015). These benefits accrue in four cardinal ways. Robot-devised recommendations of healthcare treatments or lesson plans would be better quality than those made by humans, since robots can synthesise knowledge more rapidly and are not susceptible to human error, thus leading to a more capable human workforce. Likewise, corporate strategies devised by machines will allocate factors of production more efficiently than even the best management consultants or chief executive officers can, and the implementation of artificial intelligence in research and development departments will lead to product and process innovation, which will enlarge profit margins. Additionally, since automation reduces prices for consumers and maximises revenue for firms, governments will be able to collect more taxes from both economic agents, which can fund public services crucial to productivity, notably infrastructure. The introduction of redistribution schemes, chiefly universal basic income (a program which unconditionally distributes money to all citizens), is also probable owing to its support from Silicon Valley pundits and mainstream left-wing politicians, and they will allow the poorest individuals to access basic amenities that improve their wellbeing. Such improvements would augment corporate turnover, and since workers will know that their employers have a larger capacity to pay them, they will demand more in exchange for staying with the company. This may be complemented by businesses increasing salaries as well, to make their employees happier and higher-yielding.
However, these mammoth opportunities are worthless if workers cannot capitalise on them, and although economies have myopically failed to recognise this since the Industrial Revolution, governments have finally learnt from their mistakes. They must first ensure that workers’ abilities match up to their new tasks, which is feasible because some of the necessary reskilling programs already exist, and more will surface as retraining is being taken seriously, with two-thirds of businesses reporting that it is at least a top-ten priority for them (Illanes and al., 2018). Several of these programs are sponsored by governments and can be expanded using the aforementioned tax revenue, either by financially rewarding firms who do so (e.g., the Apprenticeship Levy in the UK), or by offering them directly (e.g., Trade Adjustment Assistance in America). Moreover, technology has facilitated the birth of multitudinous online training programs, many of which are gratis, including Coursera and Khan Academy. And even these will only be needed temporarily because vocational academies and universities will soon adapt their curricula, such as by emphasising STEM subjects more heavily, to equip new generations with the skills necessary for this new work climate from the outset of their careers. Although these stints of vocational training will not completely make up for the lost talents which people spent their lifetimes building, the disparity will be minimal, as it merely takes six months to master a new skill (Gallo, 2012). Conversely, the paramount indicator that reskilling will be successful is that business executives are primarily looking for soft skills, such as communication and resilience, and these will not require significant additional resources to teach as they are encouraged by teachers and parents since childhood. The second condition for accessing these jobs is the elimination of geographical barriers, which will not be challenging given that many of these face-to-face roles have moved online; only one in five people still travel to the office full-time. In addition, many governments currently prioritise the construction of housing and transport links, which will alleviate the burden of commuting to or relocating near one’s new workplace.
It is also worth noting that automation is progressing more sluggishly than anticipated by doomsayers, which gives economies valuable time to put these structural reforms into place. The $25bn spent on industrial robots in 2020 was a hundredth of total capital expenditure, and this tentative approach towards robotisation will be bolstered by continuing calls from technology leaders, including Elon Musk and Steve Wozniak, to do so (The Economist, 2023). The adaptable nature of work in the 21stcentury will also mitigate the setbacks which will occur in the rare instances where it is impossible to access these new occupations. Industry barons such as Larry Page have suggested that workers should share jobs, instead of laying them off, which would oblige the company to pay a portion of their wages and the government to partially compensate the remainder (Lee, 2018). Furthermore, granting more leisure time to the populace will inspire the invention of new roles which were previously unimaginable; for example, the advent of the internet created a space for videogame designers and cybersecurity specialists. Governments may also now spend more heavily on hiring inside the public sector and raising wages for employees under their jurisdiction, which account for a fifth of the total workforce in the OECD (a club of mostly rich countries), and they are incentivised to do this as it would improve their favourability with those individuals (OECD, 2015).
Automation does have one momentous drawback: it will drastically reduce wage premiums for knowledge workers, including doctors and lawyers. This is because their encyclopaedic knowledge will no longer be as sought after by firms as before, since robots will offer better expertise for lower prices and erode the perks of having comparative advantage. Technology has instigated similar wage depressions in the past; pilots who flew at night used to receive 30% more pay than their counterparts who did not, as the former needed to have exceptional vision and self-control, but this is no longer true owing to better air traffic-control systems and cockpit displays (Nunes, 2021). Conversely, this decline in wages only applies to a subset of the global workforce (there are approximately one billion knowledge workers), whereas the other factors which raise salaries will be universal (Wonder, 2021).
As with the emergence of the microprocessor in the 1980s and the lightbulb at the turn of the 20th century, the techno-pessimists are likely to be resoundingly disproved about their case against robotisation. However, governments and corporations should moderate the pace of this transition and support the world’s citizens through it by giving them the opportunities to reintegrate themselves into the workforce, or they risk giving the naysayers an excuse to obstruct the benevolent march of scientific advancement.
This post is part of a debate on the effects of robots on wages. Click here for the other side.