Every Artificial Intelligence researcher has probably a vision of how intelligent machines will reshape the world. My idealistic view is one in which automation maximizes people’s freedom by giving us more time to concentrate on the things we enjoy the most. Research in all areas is astonishing: we live better and longer, and our planet has started to heal. Robots lead to superlative boosts in productivity, and an unprecedented abundance reduces tensions amongst people. We have more time to learn and inform us, and machines enable us to manage much more knowledge, so we take better decisions.
This romantic picture may be regarded as overly naive as, some would say, it disregards the immediate social implications of AI achievements. As AI becomes more pervasive, a debate around a scenario in which machines scrap people jobs- high and low skilled- is gaining steam. The discussion centers on the social disruptions generated by robots taking a prevalent role in the production system.
So far, economists have assumed that human effort will always be required. Technological change just shifts low productivity workers (more likely to be replaced by technology) into more productive tasks, increasing everyone’s well-being. Technology augments human capabilities and creates more skilled jobs. Even though specific positions disappear, people can continue working alongside machines in new roles. In this perspective, the debate around technology focuses on the education and training of the labor force.
However, critics (from left and right) are questioning whether this assumption still holds true. They believe to be supported by raw statistics. The consultancy McKinsey estimates that, as for today, around 45% of human jobs could be automated in the service sector. In the financial sector only, 50% of the work time is spent on processing and collecting data, activities that AI systems can do quite well. In manufacturing, the perspective is even more dramatic: Automation accounts for almost 90% of the job losses in the US between 2000 and 2010. Experts agree that as AI gets more sophisticated more skilled jobs across all sectors will be eliminated.
Fear of machines is certainly not new. At the wake of the first industrial revolution Luddites destroyed machines to preserve jobs. Nowadays, those worried about robots making humans dispensable are proposing (fortunately) more sophisticated approaches. Even high executives of billion-dollar tech giants- unlikely guests of the Fourth International-, are talking about socialist-friendly policies such as a universal basic income to limit the effects of automation.
The idea behind this thinking is that if we reach such a developed state of AI with little or no need for human labor, then society will divide itself between few machine owners and the rest. Most of the people without owning a robot will lack any income to buy what robots produce. At least at the surface level, this idea appears to have some commonalities with Marx’s view of the evolution of capitalism.
Marx idea of socialism as the next stage of capitalism could be (with apologies to experts) oversimplified in the following way: Capitalism leads to a natural concentration of wealth (mostly via technological change), where few capitalist get increasingly more, and a majority of workers increasingly less. The later will eventually rebel and establish a socialist system.
However, things did not happen exactly like that. Even though income inequality might be increasing, technological change led to sharp boosts in productivity, allowing everyone to be better. Marx’s workers might be relatively worst (they take a smaller share), but they also enjoy higher life standards with more goods and services available. No worth the effort of rebelling.
But what would happen if instead of making a worker produce more, machines completely replace that worker? Depriving him of any income could make the option of rebelling more appealing. Ironically, this situation might also be the end of Marxism since human labor will not be generating value anymore (apologies with experts, again!). If machines do all the work, there is no theory exploitation, and Marxism might cease to exist (it seems that even Marxist theoreticians may lose their jobs).
Marxist or not, the prospect of unhappy masses with no means of subsistence might be enough to generate a general outcry to establish some sort of direct wealth redistribution.
The proposal of a basic universal income is gaining support among businesspeople and politicians. Benoît Hamon, the French presidential candidate of the, currently in power, socialist party has proposed to finance the initiative by taxing robots. If he succeeds, every French citizen will perceive 750 euros every month no matter what, which will be, in principle, funded by robot owners.
Also from the not so socialist side, top executives from multi-billion dollar tech companies are considering the idea. They agree that the impact of automation is beyond anticipated and its consequences are getting difficult to measure. Ginni Rometty, CEO of IBM, said in Davos that “there is not one more important [topic] for all of us” than technology creating inequality and concentrating massive wealth in just a few people. She even called for an ethical rule book for AI. In Silicon Valley, the idea of “universal basic income” does not seem to be marginal. Vishal Sikka CEO from Infosys said that “a fraction of the cash that the top 10 or top 50 tech companies have could solve a big part of the basic income problem”.
A fear response or not, neo-Luddism or not, there seems to be, at the moment, a consensus that the path lead by AI won’t be that smooth and that some collective action is required. Either by direct wealth distribution, or more traditionally, via a broader and inclusive education system. Fortunately, there seems to be also a consensus that AI is a good thing and no one proposed so far to break machines (or AI researchers as in Terminator) but to alleviate the undesired consequences that technological advancements might cause.