Automation and Jobs: This Time is Different

Jihoo Lim is one of our 2020 winners for the HIEEC.


Machines, which are naturally mild, and easily kept in order, may be said now to devour men, and unpeople in times of COVID-19, as did sheep in the 15th century. The pandemic is expediting the arrival of automation which may help explain the recent K-shaped recovery. Artificial intelligence (“AI” hereafter) and robots that have been put into simple and repetitive production processes at a certain industry in the past are rapidly replacing manpower even for the tasks that were thought to be impossible only without human power. Eventually, we are threatened by concerns that AI might displace the workers permanently.

People in the past also feared that machines would steal their jobs. Ned Rudd, a young apprentice, who was believed to have led the Luddite movement, shouted to the crowd: “As more machines become available, they displace the workers, threatening our survival. Now, let's destroy those machines!” With the introduction of spinning machines into the textile manufacturing industry, which used to be made by hand, spinning machines produced much more fabric in a shorter time than the workers. As capitalists began lending spinning machines to craftsmen, workers lost their jobs, which eventually caused riots. With oppression by legal and military force, eventually the movement was suppressed. Coercive measures were not, however, the only response to the movement. For instance, concerned about the unemployment problem, the British government enacted a legislation requiring the employment of three people—a driver, an engine worker, and a rider—to operate a car. However, this legislation was eventually abolished due to its failure to keep up with the trend of advancement of technology, and the job of coachman has become extinct.


Rud and his followers have predicted that the emergence of spinning machines will permanently displace jobs. Contrary to this specter, it turns out that new jobs in many industries have been created hundreds of times as many as disappeared. We have learned from the experience of the Industrial Revolution in the 19th century that although the displacement effect may take control in the short run, the productivity effect ends up dominating, having a positive impact on employment in the longer run.


Then, will this story turn out to be the case now even 200 years after the Luddites? The scenario that rapid automation will permanently take all of workers' jobs has yet to be observed. Rather, many economists still argue that AI and robotics complement regular workers, instead of replacing them. Even job polarization emerging in recent years suggests that machines could hardly displace both high-skilled and low-skilled jobs since the former requires highly sophisticated intelligence while the latter is too economically costly to automate.


These arguments seem to assume that there are certain tasks that cannot be processed by any other entities but humans. Though this was true in the past, can such patterns thus far be the case even in the future? This time may be different from the past in that rapid progress of AI shows a possibility of a variety of high-skilled tasks displaced, including translation, academic research, medical treatment, entertainment, and other tasks requiring empathy, which have been previously considered safe from automation.


In the past, the field in which machines replaced human labor was limited to relatively mid- and low-skilled workers. By contrast, in recent years, as seen in the case of AlphaGo, technological innovation has evolved to the stage where robots can automatically learn and improve from experience without being explicitly programmed. They can, therefore, be extended to fields requiring creative or artistic sense, from doctors, lawyers, and economists to novelists and composers.


As such, it would be no exaggeration to say that we may witness that AI and robots will kill all the jobs in the present generation, opposed to what the technological innovations have done to our job thus far. Let's take self-driving automobiles for example. The invention of automobiles in the past, while displacing a coachman from history permanently, has created a large number of new jobs, including workers in automobile factories, car salesmen, car insurance salesmen, taxi or truck drivers, and even road makers, traffic policemen and junkyard employees.


Now, would the rise of self-driving cars, in turn, kill or create jobs? If AI replaces human intelligence and creativity in design, production, sales, driving, and accident handling of automobiles, all workers in automotive industries may lose jobs. Some still argue that driverless cars could create brand-new industries while endangering a great number of current jobs. We however doubt that newly created jobs are safe from attack by AI, since AI in near future are expected to perform creative or sophisticated tasks, thereby overcoming the limitations posed by those designed in the 18th century. For that reason, in 2017, Nitin Gadkari, the minister of Transportation of India, declared that it would ban the introduction of autonomous vehicles to protect jobs in the country.


What kind of future will unfold if this possibility becomes a reality? Indeed, technological progress and the limitless potential of AI can possibly boost productivity and efficiency to an immeasurable extent. Nevertheless, the wealth inequality will widen substantially provided that the fruits are distributed to capitalists who own the technology and capital. Given diminishing marginal propensity to consume, rising wealth disparity lowers the effective demand, thereby making it impossible to find someone to consume goods and services produced on an enormous scale at a lower price than in any other era, which disables economic growth from being sustainable.

Now, we note that the biggest challenges facing us going forward would be how to provide for displaced workers, and how to keep our economy sustainable in a highly unequal society. In other words, in a world where AI and robots replace human workers, the most significant task we need to address would not be how to efficiently produce, but how to fairly distribute the outputs produced by AI. Among various resolutions, it is necessary to seriously consider “universal basic income (UBI)” where the government provides every citizen with a set amount of money on a regular basis, non-proportional to the level of wealth or employment status. The intention behind this idea is to ensure that all citizens can live a minimum of human life, thereby providing them with financial security. This differs from the general social security system in that this is paid to individuals, not households regardless of whether or not they have other incomes.


However, there are a few concerns over a UBI: (1) that it could reduce incentives to work or innovate, (2) that it would not be truly effective to mitigate inequality, and (3) what could meet human desire for social cohesion and participation without jobs.


In economic terms, it is widely known that free lunch reduces economic agents’ incentives to work. Arguments for UBI are often perceived as abolishing conventional capitalistic systems as J.M. Keynes was misunderstood as a communist back in his days. However, similar to how Keynesian economics saved capitalism from the Great Depression, UBI would be a solution that enables sustainable growth by remedying workers who lose their jobs due to industrial automation.


Furthermore, we need to consider if the workers would refuse to work, satisfied with the minimum living standard. It is also crucial to ponder if imposing tax on capital income would discourage capitalists from boosting productivity through innovation.

Minna Ylikännö at the Social Insurance Institution of Finland answered this question by investigating the level of well-being and employment of UBI recipients by comparing it to a control group of ones receiving unemployment benefits. Her study concluded that UBI improves not only recipients’ financial well-being, but also their mental health, while resolving their unemployment issues as well. These results suggest that UBI does not dwindle human desire to live a prosperous life.


Next, some people remain concerned that transitioning the systematic setting of financial assistance from targeting the most vulnerable to everyone may deepen economic inequality. This seems to formulate a cogent argument. Therefore, as an alternative, we may contemplate tapered UBI, which provides zero transfers to individuals above a certain threshold while reinforcing need-based support targeting the poorest at the same time.


Lastly, in a world without jobs, in what will people find the meaning of life? There is a great concern that providing free lunch to everyone will end with destructive outcomes, such as crime, drugs, and violence. It is reasonable that elimination of concerns regarding survival weakens one’s sense of purpose, resulting in misbehavior. In order to prevent proliferation of misconduct, it is necessary to incentivize people to commit themselves to creative or artistic activities as substitutional sources of personal development. To this end, we could learn a lesson from ancient Athens, where civic class actively participated in arts, literature, and politics while leaving slaves fully in charge of production activities. As observed hitherto, it is evident that humans are likely to coexist with artificial intelligence in the near future. Despite its systematic imperfections, we note the need to assess UBI as a solution to technological advancement cutting jobs.



References


Acemoglu, Daron, “Why Universal Basic Income is a bad idea.” MarketWatch, (June 2019) https://www.marketwatch.com/story/why-universal-basic-income-is-a-bad-idea-2019-06- 19.


Agrawal, A., J. Gansm, and A. Goldfarb, Prediction Machines: The Simple Econometrics of Artificial Intelligence, Harvard Business Review Press, 272, (2018).


Archer, John E., "Chapter 4: Industrial Protest," Social unrest and popular protest in England, 1780–1840, Cambridge University