Life and work in the paradise of machines

What would we do if we didn’t have to do anything? What would a world be like where nearly all work is done by machines? If machines did all the production, humans would have to find some other way to occupy their time. They would also have to find some other way to justify the cost to society for their upkeep and their right to exist. In the current reality, one’s income is roughly tied to one’s output—though hardly in an equitable way. Investors and upper management are typically rewarded grossly more than employees for their efforts. Yet their needs as organisms are no greater. In a world where all production and most services would be done by machines, human labour would no longer be the basis for either the production or the distribution of wealth. Society would have to find some other arrangement.

In that situation, a basic income could be an unconditional human right. When automation meets all survival needs, food, housing, education and health care could be guaranteed. All goods and services necessary for living a satisfying life would be a birthright, so that no one would be obliged to work in order to live. Time and effort would be discretionary and uncoupled from survival. What to do with one’s time would not be driven by economic need but by creative vision. Thus, the challenge to achieve freedom from toil cannot be separated from the problem of how to distribute wealth, which we already face. Nor can it be separated from the question of what to do with free time, which in turn cannot be separated from how we view the purpose of life.

As biological creatures, our existence is beholden to natural laws and biological necessities. We need food, shelter and clothing and must act to provide for these needs. A minimal definition of work is what must be done to sustain life. The hand-to-mouth subsistence of pre-industrial societies involved a relatively direct relationship between personal effort and survival. Industrial society organizes production by divisions of labour, providing an alternative concept of work with a less direct relationship. Production involves the cooperation of many people, among whom the resulting wealth must somehow be divided up. Work takes on a different meaning as the justification for one’s slice of the economic pie. It is less about production, per se, than the rationale for consumption: a symbolic dues paid to merit one’s keep and secure one’s place on the planet.

Early predictions that machines would create massive unemployment have not materialized. Nor have predictions that people would work far less because of automation. Instead, new forms of employment have replaced the older ones now automated, with people typically working longer hours. Whether or not these forms of work really add to the general wealth and welfare, they serve to justify the incomes of new types of workers. As society adjusts to automation, wealth is redistributed accordingly, though not equitably. Work is redefined but not reduced. In the present economy, those who own the means of production benefit most and control society, in contrast to those who perform labour. When machines are both the means of production and labour combined, how will ownership be distributed? What would be the relationship between, for example, 99% of people unemployed and the 1% who own the machines?

With advances in AI, newly automated tasks continue to encroach on human employment. In principle, any conceivable activity can be automated; and any role in the economy can be taken over by machines—even war, government, and the management of society. We are talking, of course, about superintelligent machines that are better than humans at most, if not all, tasks. But better how, according to which values? If we entrust machines to implement human goals efficiently, why not entrust them to set the goals as well? Why not let them decide what is best for us and sit back to let them provide it? On the one hand, that seems like a timeless dream come true, freedom from drudgery at last. Because physical labour is tiring and wears on the body, we may at least prefer mental to physical activity. The trend has been to become more sedentary, as machines take over grunt work and as forms of work evolve that are less physical and more mental. White-collar work is preferred to blue-collar or no-collar, and rewarded accordingly. Yet work is still tied to survival.

Humans have always struggled against the limitations of the body, the dictates of biology and physics, the restrictions imposed by nature. In particular, that means freedom from the work required to maintain life. In Christian culture, work was a punishment for original sin: the physical pain attending the sweat of the brow and the labour of childbirth alike. Work has had a redeeming quality, as an expiation or spiritual cleanse. The goal of our rebellion against the natural condition is return to paradise, freedom again from painful labour or any effort deemed unpleasant. Our very idea of progress implies the increase of leisure, if not immediately then in the long term: work now for a better future. This has guided the sort of work that people undertake, resulting in the achievements of technology, including artificial intelligence. Humans first eased their toil by forcing it upon animals and other humans they enslaved. Machines now solve that moral dilemma by performing tasks we find burdensome. So far at least, they do not tire, or suffer, or rebel against their slavery.

On the other hand, humans have also always been creative and playful, pursuing activity outside the mandate of Freud’s reality principle and the logic of delayed gratification. We find direct satisfaction in accomplishment of any sort. We deliberately strain the body in exercise and sport, climb mountains for recreation, push our physical limits. We seek freedom from necessity, not from all activity or effort. We covet the leisure to do as we please, what we freely decide upon. In an ideal world, then, work is redefined as some form of play or gratuitous activity, liberated from economic necessity. There have always existed non-utilitarian forms of production, such as music, art, dance, hobbies, and much of academic study. Though not directly related to survival, these have always managed to find an economic justification. When machines supply our basic needs, everyone could have the time for pursuits that are neither utilitarian nor economic.

Ironically, some people now express their creativity by trying to automate creativity itself: writing programs to do art, compose music, play games, etc. No doubt there are already robots that can dance. While AI tools and “expert” programs assist scientists with data analysis, so far there are no artificial scientists or business magnates. Yet, probably anything that humans do machines will eventually do at least as well. The advance of AI seems inevitable in part because some people are determined to duplicate every natural human function artificially through technology. There is an economic incentive, to be sure, yet there is also a drive to push AI to ever further heights purely for the creative challenge and the accomplishment. Because this drive often goes unrecognized even by those involved, it is especially crucial to harness it to an ideal social vision if humanity is to have a meaningful future. Where is the reasonable limit to what should be automated? If the human goal is not simply relief from drudgery, but that machines should ultimately do everything for us, does that not imply that we consider all activity onerous? What, then, would be the point of our existence? Are we here just to consume experience, or are we not by nature doers as well?

Some visionaries think that machines should displace human beings, who have outlived their role at the top of an evolutionary ladder. They view the human form as a catalyst for machine intelligence. However, that post-humanist dream is quintessentially a humanist ideal, invoking transcendence of biological limits. It is a future envisioned not by machines or cyborgs but by conventional human beings alive today. To fulfill it, AI would have to embody current human nature and values in many ways—not least by being conscious. Essentially, we are looking to AI for perfection of ourselves—to become or give birth to the gods we have idolized. But AI could only be conscious if it is effectively an artificial organism, vulnerable and limited in some of the ways we are, even if not in all. To create insentient superintelligence merely for its own sake (rather than its usefulness to us) makes no human sense. Art for art’s sake may make sense, but not automation for automation’s sake. Nor can the goal be to render us inactive, relieved even of creative effort. We must come to understand clearly what we expect from machines—and what we desire for ourselves.