Climate change is not the only civilization threatening problem of the 21st century. In this essay I want to draw attention to the next big issue: the permanent replacements of human labor by robots and artificial intelligence. After discussing the scope and potential impact on society of this problem, I will offer a brief overview of what paths remain to prevent, mediate or compensate these risks.
The replacement of human (and animal) labor by machines is not new. But robots are not just advanced machines, just like artificial intelligence is not better software. Conceptually, a robot is a machine designed specifically to perform human capacities, and is able to do so with a certain degree of autonomy. The current wave of robotization, sometimes referred to as the ‘Fourth Industrial Revolution’, is about increasingly cheap and advanced robots that drive, sell, farm, build, talk and analyze. The latter two are essential, since automatization also threatens many service and knowledge jobs. Lawyers and financial analysts, for example. From this follows the central philosophical problem that we became advanced enough to create artificial entities that are better than us at what we do best. A tragic example of the difference between wisdom and intelligence.
Of the many ethical and social concerns surrounding robotization, the risk of mass technological unemployment is the most tangible one. But how big is this risk? Various estimates exist for Western countries. In 2013 Frey and Osborne estimated that 47% of jobs in the United States have a high right (+70%) of being automatized (footnote 1). At the time this was in line with the estimates of consultancy firms and investment banks on both sides of the Atlantic. Conservative estimates in 2016 predicted 9% of American had a very high risk (>70%) of being automatized, although many more carry some risk (footnote 2). Since these are older and focused only on United States, it might be best to look at the 2018 estimates of the OECD for 32 industrialized countries (sadly excluding Mexico). This report (footnote 3) estimates that the median job has a 48% chance of being automatized. Or alternatively: 14% of jobs are at a very high risk, while 32% have a more-than-half risk.
Two additional remarks: first, if we look at high risk sectors, this includes usual suspects such as mining, manufacturing or agriculture, but also services such as food preparation (nr.1 risk), cleaning (Nr.2) and transportation. Even jobs in sectors that carry the lowest risk, such as education, still carry a more than one in four chance of being automatized. Second remark: this risk is spread uneven across countries, and although no figures for Mexico exist, it shares risk factors with those countries at the wrong end of the spectrum, such as low educational attainment.
I am well aware of sounding like the prophet of doom, but this is a most serious matter. It is also not a ‘leftist’ topic, since I have seen surprisingly little resistance to discussing this topic with the capitalist class. It are rather their economist supporters, the media and – of course – technology engineers who have downplayed the problem. Billionaires like Bill Gates and Elon Musk have expressed concern about the issue – probably because they know what the plan is. Within Mexico conservative candidate Ricardo Anaya was the first to bring up this topic. While the debate on the extent of this problem deserves an essay of its own, optimists in general fail to grasp one of the following three things:
A. The breath of robotization, stereotyping it as a further transition from manual to intellectual labor, while artificial intelligence is exactly about replacing the second.
B. The fact that robots are aimed at replacing us, rather than improving our productivity. The self-driving car does not ‘assist’ the taxi driver in becoming more productive like when the car improved the carriage, it replaces her. Where exactly do all the cooks and waiters in McDonalds go?
C. Not seeing the latter stems from utopian thinking that disregards the capitalist context. If robots are developed with the explicit intend to replace workers and reduce labor costs, why assume other outcomes? Maybe it is feasible in an alternative context, but this is not a democratic process.
What is at risk? The capitalist social contract, which is built on wage labor. Without jobs nor the means of production, workers become extremely vulnerable to structural poverty. This both generates crisis in family income, but can also result in loss of social status and overall purpose. Making things is an important part of self-realization and work is a human right, without which humans feel incomplete. But wait, you say your daddy is rich and you intended on chilling by the pool anyway, so why care? Remember those concerned billionaires? Here’s why: if billions of people around the world lose their income, who will buy your dad’s products? Crisis will follow since one the one hand robots will allow for even more overproduction, while on the other hand reducing demand. As the early capitalist theorist David Ricardo already understood in 1821 (footnote 4): capitalism wasn’t built for this.
Like Greta Thunberg on the topic of global warming, I want you to be panicked, not hopeful. Yet not paralyzed either. Part of the problem is exactly the assumption that technological change just ‘happens’ to us, as if politics has no role to play. What is to be done? In my research I have laid out three broad courses of action: prevention, mediation and compensation.
- Prevention. Avoiding the replacement of human labor through the containment of robotization is by far the most effective and clear solution. Sin embargo, in my experience this is the path nobody involved in this field wants to talk about. Still, technology is socially constructed: it is the outcome of social processes and human decisions, which could change. If Europe can ban GMO’s and China controls the internet, robotization is a choice. In practice this dissolves into policies such as taxing robots, stopping investment in development of A.I., and subsidizing human-labor-intense industries. Tourism, for example. And on an individual level: a ‘human-made’ consumer movement rivalling the vegan one.
- Mediation. The replacement of human work does not fully translate into unemployment, for various factors ‘mediate’ the effects on the labor market. A very mainstream solution would be to massively invest in training and especially retaining, which would the transition into other jobs – if those exist. More daring would be to better spread out the existing jobs over the population through work time reduction and wage guarantees. Now many people – including tu servidor – juggle multiple jobs and/or work overtime. If we could work less, the same amount of jobs could be spread over more people. A similar effect could be achieved by allowing people to leave the labor market easier without fear of starvation, for example by increasing pensions (and/or lowing the age). For my feminist crowd: how about increased maternity or paternity leave? A truly radical option would be to question private ownership itself and support a more cooperative economy. The problem with robots in capitalism is not just that their replace workers, but that workers have no control over this process. If workers have democratic control in companies, they would be able to better guide this process or at a minimum avoid being excluded from the gains in productivity.
- Compensation. If the powers that be insist on rendering humans obsolete, they should at least financially compensate us and avoid mass poverty and starvation. When it comes to compensation, discussions usually resolve around two models. The first is Universal Basic Income (UBI), which implies periodic, typically low fixed payments (‘basic income’) to all citizens, forever (‘universal’). They are paid regardless of occupational status (Carlos Slim would get them), and are in most models – there are many – just high enough to avoid starvation. Given the enormous cost of this program, this is often funded by eliminating all other social policy, as was for example the case in the proposal of Ricardo Anaya in the 2016 election. The alternative is the expansion or (in Mexico’s case) introduction of unemployment insurance, which is aimed at protecting only those suffering from unemployment. Since this program is selective instead of universal, it could support higher payments. But unlike basic income, only for a limited duration.
Any serious political effort to confront robotization at least involves a mixture of the above approaches, since no single action (besides containment) will solve the problem. We should be wary of politicians who promise us ‘catch-all’ solutions, which only serve as an excuse to further robotize. This is often the case with Universal basic Income, which does injustice to the idea by making it more than it is. But before all of this, we must raise awareness of this new societal issue, and start forming broad alliances for the protection of human work.
So, is robotization really worse than climate change? It depends. The impact and scope of automation is clearly more limited than that of global warming. At worst it ends civilization, not life on earth itself. But from the perspective of ordinary citizens, this disaster is coming at us at a much faster phase.
(1) The Future of Employment: How Susceptible are Jobs to Computerisation? Available at: http://www.oxfordmartin.ox.ac.uk/publications/view/1314
(2) Arntz et al. (2016). “The risk of automation for jobs in OECD countries: A comparative analysis”, OECD Social, Employment, and Migration Working Papers, no. 189.
(3) Nedelkoska & Quintini (2018) “Automation, skills use and training”, OECD Social, Employment and Migration Working Papers, no. 202.
(4) Despite being an advocate of capitalism, In “On the Principles of Political Economy and Taxation” he noted the problems machines would cause for societal distribution, decades before Marx.