Published on 28 September 2018 by Soulier Avocats

Labor law and the challenges of Artificial Intelligence : 2nd part of a trilogy

Digital technology has already changed working methods. With the advent of Artificial Intelligence (“AI”), we are just at the beginning of a unparallel transformation that will affect not only the labor and employment market but also working relationships. What does exactly mean AI’s impact on working relationships? When we say working relationships, it implies labor law.

Labor and employment law should be used as a legal tool to steer the obvious changes brought by AI in the workplace. The challenge is thus to identify avenues for adapting our labor and employment legislation in order to anticipate and smooth the transition to the new world.  

This article is the second part of a trilogy built around the lifetime of employment contracts:  hiring / performance / termination. It is devoted to three issues connected with the performance of the employment contract: What tools can be used to ensure the continuing adaptation of employees to their jobs that are evolving with the development of AI, what is the impact of AI on the ways of organizing work within businesses, what is the impact of AI on employees’ working conditions?

In the first part of this trilogy, I addressed issues related to the termination of the employment contract and I specifically wondered whether our labor and employment legislation, as it currently stands, provides some safeguards against the unavoidable (according to some people) risk of job elimination as a result of the development of AI?

While technological changes constitute today a ground for dismissal per se, it seems reasonable to assume that, when it comes to assessing whether the dismissal on economic ground due to technical changes – e.g. the introduction of an AI-related tool – is justified, judges of tomorrow will carefully look at the efforts deployed by companies to adapt their employees.

Consequently, it is necessary to anticipate changes in jobs and skills in each industry to avoid large-scale lay-off plans. Continuing training is a key challenge of the digital revolution.

This brings me to the second part of the trilogy on employment contracts lifetime: the performance stage.



In this second part, I will try to answer three key questions:

  • Which legal tools that already exist are available to companies and employees to ensure the necessary constant adaptation of employees’ skills?
  • What is the anticipated impact that AI will have on the ways of organizing work?
  • What is the anticipated impact that AI will have on working conditions?
  1. Continuing adaptation/training of employees: What legal tools?

Businesses must make the most of the digital revolution by defining an appropriate human resource management policy.

We can indeed expect that the actual impact that the development of robotics and AI will have on employment will also depend on the ability of employees to master news tasks and to gain new skills in order to bring an “added value” while competing with increasingly efficient machines.

As such, do we have today the legal tools that are necessary to start the adjustment of businesses to AI and ensure the adaptation of employees to their employment position?

The answer is yes and the most efficient legal tool is indisputably the Gestion prévisionnelle de l’Emploi et des Compétences (forward-looking job and skill management policy, or hereinafter “GPEC”).

The GPEC is an essential component of human resources management policies within companies. It is about devising, implementing and monitoring coherent policies and action plans:

  • To reduce at an early stage the gaps between a company’s needs and its human resources (in terms of staffing and skills) by anticipating the foreseeable evolutions of jobs, skills and qualifications in light of foreseeable economic, demographic and technological changes,
  • To involve employees in a career development project.

A. Existing mandatory forward-looking job management schemes

Several provisions impose on employers the obligation to take measures that fall within the scope of forward-looking job management policies:

  • Generally speaking, the employer has the obligation to ensure the adaptation of its employees to their jobs. According to Article L.6321-1 of the Labor Code, “the employer must ensure the suitability of employees for their post and oversee their capacity to perform their job, particularly with regard to employment, technological and organizational evolutions”.
  • Forward looking job management agreements are submitted to the Social and Economic committee (“SEC”) as part of the information/consultation process. Indeed, the SEC must be consulted on “the company’s strategic directions, as defined by the company’s management or supervisory body, on their consequences on the business activity, employment, the evolution of employment positions and skills, work organization, the use of subcontractors, temporary workers, fixed-term contracts and internships. This consultation process also includes the forward-looking job and skill management and the orientation of vocational training. (Article L. 2312-24 of the French Labor Code).

The two aforementioned legal provisions indirectly impose on employers the obligation to define a GPEC.

  • In addition, pursuant to Article L. 6315-1 of the French Labor Code “at the time of hiring, the employee is informed that he/she will benefit from a career development review meeting with his/her employer every two years. This review will address the employee’s career development prospects, especially in terms of professional qualifications and employment. The purpose of this review meeting is not to assess the work and performance. It must also address information on the validation of prior experience.” Every six years, the career development review meeting shall include a status report on the employee’s career path within the company.

This measure providing for a review meeting contributes to the management of jobs as it deals with the “employability” of employees.

  • Lastly, some provisions encourage employers to negotiate in order to conclude a GPEC agreement or to define the contents of such agreement: Pursuant to Article L. 2242-2 of the French Labor Code, every three years companies of a certain size must launch negotiations on the management of jobs and professional career paths as well as on gender diversity in the workplace which must address the GPEC policy.

B. The GPEC strictly speaking

Regarding the GPEC legal arsenal strictly speaking, it is only mandatory for companies with at least 300 employees:

  • Companies and groups that employ at least 300 employees have the obligation to negotiate on the GPEC at least every three years (Article L. 2242-13 of the French Labor Code);
  • For companies with less than 300 employees, there exists a so-called “support package” endorsed by the State in order to co-finance a HR diagnosis and prepare a GPEC recommendation/action plan. The State bears 50% of the costs up to a maximum of 15,000 euros.

Concretely speaking, the GPEC requires the implementation of various types of measures, such as monitoring the evolution of employees, review meetings, skill assessments, career advancement processes, professional development paths (“cold” management of jobs), professional and geographical mobility, etc. If jobs are threatened, the GPEC can also be a means of collective termination of employment. Training and mobility are the two keywords of the GPEC.

Some of the aforementioned topics are imposed by law and must be addressed during negotiations (Articles L. 2242-20 and L. 2242-21 of the French Labor Code).

It is interesting to note that “the qualification of categories of jobs threatened by economic and technological developments” is among the optional topics of negotiation. While this item, which is not mandatory on the date hereof, may appear to be more a question of prospective analysis than negotiation, the odds are that it will become a key essential topic with the advent of artificial intelligence.

Let us take a very concrete example. Here is what the GPEC can do:

  • A diagnosis: The job of checkout operator is threatened by technological developments;
  • An action: Migration of a declining job – checkout operator – towards an emerging job: Supervisor of self-checkout machines;
  • An accompanying measure: Theoretical and practical training courses.

Of course, even though the legal arsenal already exists, it is absolutely necessary that employers and employees play by the rules, i.e. employers must negotiate fairly and implement the negotiated action plans on the one hand, and employees must accept to evolve and to follow training courses…

C. Assessment/prospects: Axes for improvement suggested in the Villani report

In his report entitled “For a meaningful Artificial Intelligence” released on March 28, 2018, Cédric Villani suggests amending the contents of both the mandatory annual negotiation and the triennial negotiation on the GPEC in order to factor in the introduction of new technology and the digital transformation of companies, in terms of adapting skills and of complementarity between humans and machines.

This is precisely the challenge: Adapting today’s workers and preparing them for the jobs of tomorrow.


  1. Should we rethink the ways of organizing work in the age of AI?

A. Work organizations are already evolving (telework, etc.)

Things have already started to change with the development of digital technologies that have radically affected the space/time organization of work: Telework, co-working, etc.

 Hélène Carrère d’Encausse, historian and academician, wrote “Space and time had always dominated the lives of men; all of a sudden, they can master them both through progress. Everything happens here and immediately”.

In addition, the new working world where AI is present imposes to rethink the relationships between employees and machines. Machines no longer serve as a working tool but as a support for employees or, even in some cases, they replace him/her. Examples include the so-called “LegalTech”, i.e. these start-ups that offer legal services, settle minor disputes or predict the chances of success of a case. 

Yet, robotics and AI relieve men of repetitive tasks and thus have the benefit of transferring men’s creativity to high added-value activities.

As such, according to Cédric Villani in his aforementioned report[1], “The automation of tasks and occupations could represent a historic opportunity for de-automating human work: It enables us to hone our uniquely human skills (creativity, manual dexterity, abstract thinking, problem-solving)”. This is what Cédric Villani calls “enabling complementarity”.

B. What types of work organizations could be expected tomorrow?

 Prospectively speaking, we can expect a more flexible exercise of professional activities, with a way of working more and more in a collaborative manner.

  • Management: New work organizations should be less pyramidal. The manager shall become more of a coordinator than a supervisor. His/her technical skills would not matter as much as his/her ability to moderate and lead a community and master the communication channels. In his report entitled “Transformation numérique et vie au travail”, (literally “Digital transformation and life in the workplace”, Bruno Mettling explains that “this brings about a change of leadership mode that used to be based on the holding of information and that will be based in the future on the ability to reach consensus.
  • Special attention should be paid to the danger of decision-making support machines: There is a human tendency to have greater confidence in the solution suggested by a machine. The danger could hence be to transfer decision-making responsibility to machines. Indeed, we all known that machines also can make mistakes.

C. Assessment/prospects: The current change in work organizations must be discussed within businesses

 The necessity to regulate new work situations requires the launch of a specific legislative project. Labor law, which still primarily relies on the unity of time and place of work, must adapt accordingly. 

 Cédric Villani recommends that “an enabling form of complementarity develops within organizations”. According to him, this form of complementarity must be broadly discussed, first and foremost among employees. Indeed, the aim is to encourage employees to develop their creativity while protecting those for whom calls for creativity can be problematic.

In businesses, it is necessary to anticipate and to design training plans for tomorrow’s managers.


  1. How to better regulate “new work situations” to limit the emergence of new psychosocial risks?

A. Regarding working conditions, labor law, which is principally geared for working methods developed during the industrial era, is insufficient

As indicated by Cédric Villani in his aforementioned report[2], complementarity between human and machines raises the question of what framework should govern working conditions in the digital age.

Is our current law that regulates working conditions adapted to the development of AI in everyday business life?

The answer is no: The legislation governing working conditions is primarily geared for working methods developed during the industrial era.

There exist some areas of progress, in particular telework and the right to disconnect. The recognition of a right to disconnect enshrined in Law of August 8, 2016 is part of the efforts to control the impact of digital devices on employees’ private life and health.

However, it seems that a series of new risks and situations, related to the development of digital technologies in companies, has not be sufficiently considered.

 As a result, for a large number of businesses and individuals, the framework governing working conditions is too rigid and, in some cases, simply unsuited to new working arrangements, and it does not take the specific new risks into consideration.

It is true that robots can take on tedious tasks and avoid that humans be exposed to dangerous conditions, such as during the depollution of sites contaminated by toxic waste.

As such, robots would help reduce physical arduousness. But there are other types of arduousness at work, aren’t there?

B. What are the new risks?

  • Loss of the ability to take initiatives when men only follow the instructions of a robot

The word “de-automation” of human tasks has been used. However, this is only the case where machines completely replace humans who can then engage in other tasks with high added-value. Yet, another form of work can exist: A collaborative work with a machine. Major retail logistics warehouses provide a typical example of this ambiguity: The automation of processes may lead to employees solely following orders from a machine (“voice picking”).

In his aforementioned report[3], Cédric Villani points out that “Working in collaboration with a machine may increase the routine nature of tasks and reduce capacity for personal initiative and thinking, at times under the guise of improving working conditions. Although a certain form of automation may evidently make life easier for employees, the longer-term risks nevertheless need highlighting.”

Indeed, the automation of processes may lead to employees solely following orders from a machine.

Following orders from artificial intelligence, losing control over processes, delegating decisions to machines are all examples of complementary working which, at individual and collective levels alike, (…) may result in suffering at work.

  • Increased working hours since robots are not subject to working time regulation

The fact that robots will be handling tasks currently subcontracted/outsourced in developing countries can paradoxically entail a risk of increased working hours for the machine supervisor and thus create another source of stress at work.

  • Isolation in the face of machines

We can also imagine another sort of arduous working conditions linked to employees’ isolation in the face of machines: Exclusively following a machine’s instructions, no possibility of discussing with colleagues without going through a machine interface, etc.

 Are we moving towards a new form of arduousness at work? But the so-called Compte Pénibilité account[4] has been structured around the concept of arduousness as it existed during the industrial age.

C. Assessment/Prospects: Avenues for reflection on the prevention of tomorrow’s psychosocial risks

The reflection concerns the measurement of the workload since working time duration is no longer sufficient to assess the workload. The keyword here is “feelings” because we are talking about emotion and perception, this “little extra” that still enables us to distinguish ourselves from AI-powered machines.

Several proposals have already been put forward:

  • The Anact[5] suggests putting in place activity indicators to assess the adequacy between the prescribed workload, the actual workload and the perceived workload (i.e. the workload perceived by the employee);
  • Bruno Mettling’s report “Digital transformation and life in the workplace” released in September 2015 recommends “making workload an item of discussion, regularly addressed at the level of the team, not at the level of the individual employee, it being the manager’s duty to subsequently consider the feelings of employees. The manager’s role is indeed decisive to prevent psychosocial risks insofar as the virtualization of work relationships can reduce the “listening time” and the time dedicated to the coordination of actions. The training of line managers thus becomes a necessity, in particular to avoid harassment through deregulated remote management methods”.

This brings me to the third part of the trilogy on employment contract lifetime: The conclusion of the contract, i.e. the hiring of employees.

For employees to be hired, there must be vacant jobs and “employable” employees!

As indicated by Yuval Noah Harari in his book “Homo Deus: A Brief History of Tomorrow:

The crucial problem isn’t creating new jobs. The crucial problem is creating new jobs that humans perform better than algorithms”.

To be continued…


[1] For a meaningful Artificial Intelligence, report authored by Cédric Villani and released on March 28, 2018

[2] For a meaningful Artificial Intelligence, report authored by Cédric Villani and released on March 28, 2018

[3] For a meaningful Artificial Intelligence, report authored by Cédric Villani and released on March 28, 2018

[4] This device is designed to improve the status of workers subject to arduous working conditions either by allowing lighter work schedules or job reclassification or early retirement

[5] Agence nationale pour l’amélioration des conditions de travail, i.e. French Agency for the Improvement of Working Conditions