Three myths about the future of work according to Daniel Susskind.

In his TED talk delivered in Germany in December 2017, Daniel Susskind, a professor and researcher in economics at King’s College London, offers a reflection on the profound transformations that robotics and artificial intelligence will bring to bear on work, employment, and society. The author of several books, including A World Without Work (Flammarion, 2023), in which he considers the end of work and the advent of a state-regulated leisure society, Susskind more broadly analyzes the possible impacts of the development of AI technologies. In his view, the central question is not only whether new technologies pose a real threat to employment — which he accepts — but rather what, in the future, will be the unprecedented problem that humanity will have to solve. To shed light on this debate, he argues that our vision of automation remains clouded by three myths that must be dismantled: the Terminator myth, the intelligence myth, and the superiority myth.

• The Terminator myth: We must first rid ourselves of the idea of a world of work gradually and entirely replaced by robots. Certainly, many tasks once performed manually are now automated, but these technologies also enrich and increase the value of human work. They do not eliminate the activity in question; they make it more efficient. Susskind cites the example of the taxi driver using a satellite navigation system, as well as that of the architect working with computer-aided design software. This complementarity between human work and technological progress produces a double positive effect: 1) It increases overall output through productivity gains, thereby facilitating labor mobility and the retraining of affected workers. 2) It transforms demand: additional income encourages the emergence of new products, new industries, and, consequently, new tasks and new occupations. An illustration of this phenomenon can be found in the historical evolution of the distribution of employment across sectors: from the primary sector (agriculture), to the secondary sector (industry), and then to the tertiary sector (services). Over the span of three centuries, the majority of the population moved from farms to factories and then to offices. The true challenge raised by the “Terminator myth” therefore lies in the capacity to retrain workers affected by technological transformations.

• The intelligence myth: The second myth to dismantle is what Susskind calls the “intelligence myth.” Until recently, many economists believed that tasks such as driving a car or diagnosing medical conditions were difficult to automate. We now know that this is no longer the case. Many tasks requiring high-level cognitive abilities can be carried out by artificial intelligences with an efficiency sometimes superior to that of humans. According to Susskind, economists were misled by the idea that automating necessarily requires reproducing human modes of thought. In reality, machines do not copy our reasoning: they perform the same tasks differently, using models trained on massive volumes of data. What matters is therefore no longer the way in which a task is carried out, but the fact that it is carried out, even according to radically different logics. Understanding the human brain thus becomes secondary: machines act according to their own principles, without seeking to imitate human intelligence.

• The superiority myth: The third myth is that of “superiority.” Supporters of the Terminator myth assume that machines can only supplant human workers. They thus fall into what the economist D. F. Schloss already in 1891 called the “lump of labor fallacy”: the idea that there exists a fixed quantity of jobs to be distributed. Yet history shows that increased productivity instead generates more output, more demand, and, in the end, more work. Susskind does not deny that machines substitute for certain human tasks, but he warns that, in the near future, complementarity may no longer be sufficient to create new opportunities for humans. The example of driving is illuminating: today, GPS assists drivers; tomorrow, autonomous vehicles could replace the driver entirely. In that scenario, output would continue to increase, but new tasks would be performed by machines rather than by humans. Value creation would therefore no longer guarantee the creation of human jobs.

Lessons and recommendations: From his analysis of these three myths, Daniel Susskind draws several lessons:

• The Terminator myth shows that the future of work depends on the balance between substitution and complementarity.

• The intelligence myth indicates that substitution progresses continuously and that there is no solid reason to anticipate a limit to it.

• Complementarity remains beneficial as long as the new tasks generated by growth also benefit humans, and not exclusively machines.

• But the superiority myth reveals that this complementarity tends to shrink, suggesting a shift in favor of machines.

By combining these three dynamics, Susskind sketches a worrying future: increasingly capable machines, able to perform a growing number of tasks, progressively reducing the space for complementarity between humans and machines. The traditionally beneficial effects of technical progress could thus be reversed, to the almost exclusive benefit of machines. He nonetheless concludes on an optimistic note: if technological unemployment is the symptom of humanity’s technical success — marked by the spectacular increase in output since the Industrial Revolution — the real challenge is no longer to produce more wealth, but to distribute it better. In other words, the question is how to ensure that the prosperity generated by capitalism benefits everyone in a world where the role of human labor tends to decline. This is, according to Susskind, the fundamental problem that our societies will now have to face.