Published on 27 April 2018 by Soulier Avocats

Will AI spell the end of human work?

The issue of the impact of artificial intelligence (“AI”) on the workplace unleashes fierce passions: Some fear – sometimes rightly – a massive destruction of jobs in the coming years because of the emergence of AI. Others are more moderate and rather anticipate an evolution in the way we work, through the adaptation and redefinition of jobs, and the creation of new opportunities. What is the real truth?

History shows us that previous periods of transition did not always run smoothly[1].

It is true that the invention of the printing press by Gutenberg in the 15th century – 2018 marks the 550th anniversary of Gutenberg’s death – hastened, despites the bitter struggle led by master writers and illuminators who formed a guild (i.e. an association of craftspeople in a particular trade), the disappearance of the copyist’s profession[2]. Similarly, the industrial revolution in the 18th century led to the elimination of many categories of jobs, including that of “canut[3].

However, the invention of the printing press by Gutenberg also entailed the creation of the typesetting profession. The industrial revolution, “stream of innovations[4], has indisputably played a leading role in the development of new ways of working (such as Fordism[5]) and, correlatively, in the creation of new trades and new types of jobs. Schumpeter wrote “The process of industrial mutation that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one. This process of creative destruction is the essential fact about capitalism. It is what capitalism consists in and what every capitalist concern has got to live in”.[6]

The process of “creative destruction[7] should most likely apply once again with the advent of AI, thereby leading to the creation of new functions, in particular for the supervision and maintenance of AI systems.

In addition, and with all due respect to alarmists and prophets of doom and gloom, many studies show that less than 10% of jobs would be at risk: According to a recent analysis made by McKinsey & Co, fewer than 5% of occupations can be entirely automated and about 60% of occupations could have up to 30% of their constituent activities automated [8].

This is the reason why it is necessary today to redefine certain jobs/occupations, and to rethink workers’ education and training: Cédric Villani calls for “an adaptation of the initial education offer which prepares citizens for traditional jobs impacted by AI, and the creation of new paths that correspond to new jobs generated by AI”. The lifelong training offer also needs “to evolve and to expand in order to enable citizens to follow retraining programs and to upgrade their skills, or to invent new activities, in pace with the innovations induced by AI[9].

As such, we must rethink not only the substance, the content, by creating new AI-related training programs and thus “triple the number of people trained in AI in France in the next three years[10], but also the form, by further stimulating creativity, as automatable tasks are intended to be eventually fully delegated to machines.

Ultimately, it is undeniable that AI will replace humans in many areas, starting with those that included automatable tasks. Useless to try to compete: AI will always be better than us.

However, these areas are limited, quite simply because – as rightly pointed out by Jean-Gabriel Ganascia[11] – intelligence goes beyond computation capacities and the storage of enhanced information. Other more or less controversial writers endorse this approach: Howard Gardner, who is behind the theory of multiple intelligences, Antonio Damasio[12], who considers that reason and emotion are reliant on each other, and many others.

Today’s priority is thus to “develop the means for effective complementarity between human tasks and machine tasks[13].


[1] Report on Artificial Intelligence, Cédric Villani, 2018.

[2] A copyist was a person who made copies of handwritten documents before the invention of the printing press. This job was originally done by monks.

[3] Silk workers of the City of Lyon.

[4] Capitalism, socialism and democracy, Joseph Schumpeter, 1942.

[5] Standardized mass production and assembly line (cf. in particular Charlie Chaplin’s Modern Times, 1936).

[6] Cf. footnote 4.

[7] Cf. footnote 4.

[8] Four fundamentals of workplace automation, McKinsey Quarterly, 2015.

[9] Cf. footnote 1.

[10] Cf. footnote 1.

[11] Le Mythe de la Singularité, (i.e. The Myth of Singularity), Jean-Gabriel Ganascia, 2017.

[12] Descartes‘ Error, Antonio Damasio, 1994.

[13] Cf. footnote 1.