Originally published in French on Technoprog 

Of course, this terra incognita ahead of us is feeding a collective angst toward the possibility of an unfriendly AI eradicating humanity, up to the brightest minds: Jacques Attali in France, economist and social theorist, asked for a moratorium on AI research following a Google publication about an experiment on two artificial intelligences inventing encryption of their own that a third AI failed to decipher; or also the alert raised by Stephen Hawking, Bill Gates and Elon Musk (see note ). Even if we’re ruling the artificial intelligence existential risk issue out, which does deserve a very precise analysis, the less dramatic hazards of a society without jobs will also challenge us. What of a world where machines dominate all production activities without any place for human beings - outdated, too slow and lacking the endurance to sustain a task to meet the new profitability standard dictated by autonomous machines?

Robotics and artificial intelligence, like any innovation before, are enforcing on employment and profitability the sort of creative destruction theorized by Joseph Schumpeter. But from now on,  the creative action caused by innovation will no longer be good enough, with machines replacing people in both existing and newly created jobs, to the point that the economy fails to generate sufficient new activities which require a human operator to be performed.

It’s then urgent to anticipate this evolution, because from inception to conclusion, whether potential social disruptions will take place peacefully or violently, will depend on the strategies we will choose to follow.

An Old Basic Shift

The phenomenon we’re talking about has already begun. It came into being from the advent of the industrial revolution when the 19th century brought civilization into the machine age. At first, those machines, even though they enabled one worker to accomplish the work of one hundred, were still tools needing human operators.

The first step was rural flight. Agricultural mechanization triggered a massive rural exodus toward metropolitan centers where newborn industry absorbed that cheap workforce en masse. (By the way, this applies to China and its migrant workers, who have fueled the industrial rise of the country since the end of the 20th century, with all the social and human issues such a mass uprooting can induce.

Here’s our first delusion: when automation arises, there will always be a new set of activities opening up for those left behind; except in this case automation is arriving on both sides of the situation.

Then, as a precursor for a second step, Henry Ford invented fordism and its corollary, assembly lines which create a terrible sense of alienation among factory workers, a feeling brilliantly illustrated by Charlie Chaplin in “Modern Times” or Fritz Lang in “Metropolis.” But assembly lines were also at the root of a greater wealth, giving birth to a society of abundance, the consumer society. Beyond its constraints, dividing an activity into unitary actions spread along assembly line positions set the stage for automation to begin. As a logical consequence, industrial robots appeared in factories in the US as early as the 1960s before expanding worldwide in the 1970s. This robotization of industrial factories, paired with offshoring of non-automatable tasks to emerging countries with low wages, rapidly diminished the demand for industrial labor in old western industrialized economies.

However, the significant savings in production costs enabled by that double trend contributed to the expansion of consumer society, which generated new needs, and subsequently employment increased dramatically to manage product flows through international trade and commercial hubs like shopping malls. This newly available workforce was absorbed by the tertiary, or service, industry which quickly became the dominant sector in developed economies.

Here’s our second delusion: the idea that a wage-earning employment system closely linked to the capitalist age would last forever, and thus that adaptation to innovations means reinventing traditional wage-earning employment to follow. But the incoming change involving overwhelming automation of labor isn’t a simple evolution, it’s more a shift to a new era of civilization. Humankind has gone through such a radical change only twice in the past with a shift of this scale: first from the paleolithic to neolithic; and second from agrarian societies to the industrial age.

David Autor And The Rise Of Polarization

So the question everybody is asking anxiously remains an open-ended one at this point: will mass automation of the economy wipe out labor? Probably not yet. Most optimistic analysts are predicting general artificial intelligences by 2040, while most pessimistic ones predict they will take more than four thousand years. Such a huge gap is interesting because it captures the big number of unknowns in that equation, and thus demonstrates all the complexity which will need to be managed when strong AI does arise.

Even though most optimistic projections indicate a short period (around twenty years) we still have enough time left to get ready for this new era.

Facing such societal pressure, social instability will very likely dangerously escalate if nothing is done with regard to this new situation. Counting solely on Schumpeterian creative destruction to help us out of this gridlock is just another self-deception. It’s obvious that creative destruction is still a factor but alone it won’t hold up against such a tectonic shift. Nevertheless, the measures to be taken to mitigate the effects of a labor transition should not be worked out after the fact, lest they be counterproductive. We should, by our own volition, move with this shift and adapt how we work on a regular basis.

In that respect, a universal basic income seems to be one of our best tools for a medium-term solution, but also a tool which brings significant immediate risks.

From negative income tax to universal basic income

The main risks of basic income depend on the way it’ll be implemented. Implementing it too soon or too late vis-à-vis the wage-earning situation can potentially trigger toxic uncontrolled effects highlighted by some French laissez-faire liberals. They consider those drawbacks as an ontological toxicity of basic income and the roots of what they see as the ideological disaster of the century.

It’s true that basic income is an optimistic bet, and if activated too early it is undoubtedly unrealistically dangerous thinking to believe that wages for difficult unskilled jobs will naturally increase, in order to seduce back to the workplace those people previously demobilized by the security of a universal basic income. Changing minds is never that easy. If UBI is totally decorrelated from economic reality, it would be at best ineffective, and at worst harmful for economic vitality by punishing entrepreneurship and active workers to the point that everybody would be penalized, even those it is supposed to help. As a first step, with a polarized economy as David Autor described it, a negative income tax may be more effective due to the fact that the labor market, even though it has started to dry out, is still a reality. During the Swiss referendum of 2016, a universal basic income proposal was rejected. During the campaign, researchers at the Institut de Géographie et Durabilité (Institute of Geography & Sustainability) at the University of Lausanne argued that the negative income tax would be a better solution to mass technological unemployment, especially given that basic negative income taxes already exist, as in France with the RSA. However, with time Moravec’s paradox will be broken. There’s a strong chance that the DARPA Robotics Challenge will play a key role in it, having started, through its positive competitive constraints, a virtuous loop towards breakthrough innovations. For example the DARPA Grand Challenge in the early 2000s launched autonomous vehicles, a development which is in turn today revolutionizing occupations like taxidriver, whose very existence is threatened.

Once Moravec’s paradox is relegated to the history books, formerly protected unskilled low wage jobs will disappear due to automation. The only jobs to persist will be those needing creativity and an ability to work with artificial intelligences, an aptitude that will be seen as a positive skill and then as a way to achieve a larger social valorisation. At some point, the switchover to mass automation will be so overwhelming that the negative income tax will probably become toxic while a universal basic income will more efficiently stabilize our societies.

Idleness society and copyright society

To conclude, consider a hypothetical question: What will society look like at a time when the last human with a wage-earning job loses it?

Rolling out a universal basic income won’t guarantee that everything will be fine. Some fundamental issues will have to be solved: funding, which leads directly to a danger of  loss of control on expenditures required by this welfare income; and we will also have to be careful to avoid creating “annuity societies”. In the societies which favorize stability and status-quo, such creative destruction through innovation appears as a dangerous risk. Then innovation would be contained by artificial regulations in order to prevent that society to be challenged. Those kind of social structures would give birth to a world where innovation and progress would give way to a quest for a status-quo more favourable to maximizing one’s government issued annuity, than for change and disruptive technologies.

To avoid a society of inaction, the link between the real economy and income must be kept alive. Nowadays, a majority of people experience that linkage through alternate periods of prosperous full employment, and less dynamic periods of mass unemployment. This linkage also supports innovation, rewarded when audacious entrepreneurs have a good return on investment because of a well-thought-out and successful R&D strategy. For the latter,  the starting point is a sound intellectual property policy, which consistently rewards creativity and which sets an example for strong incentives to develop groundbreaking solutions.

In addition, a mix between crowdfunding, massive online open courses and pervasive 3D printing technology may lead to a post-industrial society where individuals are far more autonomous. Crowdfunding will enable huge funding out of public money or without dependency on great industrially dominant companies; MOOCs will give free access to knowledge for all, making social barriers to access to education obsolete; and generalized 3D printing will make obsolete factories, economies of scale, and flow or stock management, enabling someone to access the worldwide global market without any need of support from a major industrial group. In such a society, R&D income won’t go exclusively to large, faceless corporations, but to inventors, as with copyrights. And if copyright regulation is implemented to have an optimal duration - which is around 15 years according to a study by Rufus Pollock - thereupon we might be at the threshold of a new era of economic prosperity.

However, even in that society, we must keep universal basic income connected with the real economy by keeping it connected to GDP. With that link, basic income will diminish during times of economic crisis, forcing people to question their behaviors. Perhaps it will enable a wide understanding in the general public opinion for the need to support those innovation-friendly policies and measures increasing economic fluidity, which are favourable to full prosperity.

In the long run, it’s quite likely that “handmade work=” will take on a new meaning as result of its scarcity; feats of gifted people, athletes, artists or performers will be sought after and those persons will earn fame and better income. Other creative activities will be those exclusively done by artificial intelligences or, within the human race, by individuals who agree to be enhanced to follow the machines’ pace of work. If nothing is done though to avoid brutal downgrades for people that become unemployable due to biological limitations, we will slide into dystopian societies trapped between the disappearance of humankind, and ultraviolent revolts very disruptive for both societies and economies. Unless action is taken, society will have greater inequality than we have today. And very polarized societies are known to be generally unstable and dangerous ones, including for the wealthiest individuals.

That said, even if today a universal basic income seems unavoidable after the disappearance of work, the path to that era is still a long one. It won’t be a surprise if new, breakthrough modes of operation arise under constraint to propose better ways to organize our societies without any need of a universal basic income to keep them stable and wealthy.