Social plans justified by the productivity gains brought by AI are multiplying in the United States and now in Europe. French companies operate in a different framework.
These massive downsizings have the common point of being justified, in whole or in part, by the productivity gains brought by AI. In a letter sent to his employees, Amazon CEO Andy Jassy makes an obvious connection. The deployment of generative and agentic AI will change the way the e-commerce giant works. “We will need fewer people to do some of the current jobs, and more for other types of trades.” However, the scales are tipping on the wrong side and the manager plans, ultimately, to reduce the workforce of his company.
Consulting firms, which insist to their clients that AI does not replace humans but on the contrary augments them by relieving them of tasks without added value, are clearly not applying this strategy to themselves. CEO of Accenture, Julie Sweet has clearly explained that she wants to remove consultants who do not adapt to AI. “We separate in a very short time from people who we think, based on our experience, will not be able to learn the necessary skills,” she explained, according to comments reported by Les Echos.
The “jobs apocalypse” extends to Europe
This “jobs apocalypse” does not only concern the tech sector. The German insurer Allianz will cut, in its subsidiary Allianz Partners, up to 1,800 jobs in its call centers to replace them with AI and other chatbots. Also across the Rhine, the airline Lufthansa is banking on AI to destroy around 4,000 jobs by 2030. The Dutch bank ABN Amro plans, through its AI plan, to eliminate 5,200 jobs, or 24% of its workforce. And when AI does not replace humans, it slows down or freezes hiring.
Is this wave of social plans the prelude to even more massive destruction? For MIT researchers, the announced layoffs would only be the visible part of the iceberg. In a recent study, they estimate that current AI already has the capacity to replace 11.7% of the American workforce, or nearly 20 million employees working in administrative services or the financial function.
While robotization has depopulated factories, this new wave of automation threatens tertiary sector jobs. CEO of Anthropic, Dario Amodei suggests, according to an interview given to the Axios site, that AI could “eliminate” half of entry-level white-collar jobs and increase the unemployment rate from 10 to 20% within five years. Other voices are less alarmist. In a column published by Le Monde, a trio of economists including 2025 Nobel Prize winner Philippe Aghion, estimate that AI should only increase productivity by 0.8 to 1.3% per year over the next decade.
A protective social framework
At an individual level, Stéphanie Poussou, of counsel at Capstan Avocats, has not yet been aware of a dismissal for professional inadequacy linked to an unsuitability for AI tools. “No dispute relating to this reason has yet been examined by the Industrial Tribunal. However, it is still too early to observe the first decisions, knowing that an employee has a period of one year to contest his dismissal and that the procedures are currently long.”
Furthermore, French law imposes an obligation of adaptation and professional training. Article L6321-1 of the Labor Code provides that “the employer ensures the adaptation of employees to their workstation” and that he “ensure that their ability to hold a job is maintained, particularly with regard to the evolution of jobs, technologies and organizations.” “It can offer training that contributes to the development of skills, including digital skills (…)”
“If an employer dismisses an employee for professional inadequacy, the industrial tribunal may criticize the employer for not having supported the employee in improving their skills, through training for example,” says Stéphanie Poussou. On the other hand, an employee who is resistant to change and refuses any training and support action could be dismissed for this reason. Furthermore, French law also incorporates the principle of non-discrimination. “Targeting employees perceived as being overwhelmed by AI could, indirectly, affect certain age groups or profiles, which exposes them to a risk of litigation,” she continues.
Far from the brutal methods of their American counterparts, large French companies prefer to communicate about their program to support their employees on the path to AI. For Marc Trilling, CEO of the Saegus firm, the question facing them is not so much the possible job cuts as how to train or even retrain their employees so that they can fully exploit these new AI tools. “It is crucial to offer employees professional paths that allow them to either maintain their specialty or move to new functions. The employee retains their expertise while being augmented by AI.”
BPCE and Macif lead the way
This HR transformation project involves mapping existing skills, anticipating tomorrow’s expertise needs and then developing upskilling or reskilling actions. For Marc Trilling, the accelerated pace of AI requires companies to review their skills more frequently, every 18 or 24 months compared to the usual five to seven years. According to him, AI can be used. “By analyzing in real time interactions and developments in skills within the organization, AI will dynamically feed a skills repository, far from the traditional and fixed practices of what is often perceived as an imposed exercise.”
To carry out this work, companies can rely on, recalls Stéphanie Poussou, on a valuable tool that is Job and Career Management (GEPP), mandatory for those who employ more than 300 employees. This system makes it possible to anticipate transformations in professions, to plan suitable training systems and to reflect on professional mobility trajectories. “No negotiation or renewal of a GEPP agreement should be done without integrating a component dedicated to AI,” she believes.
This is BPCE’s bias. At the end of September, the banking group, which brings together the fourteen Banques Populaires and the fifteen Caisses d’Epargne, signed, unanimously by the representative union organizations, a GEPP agreement fully taking into account the impacts of generative AI. To bring its more than 100,000 employees on board, the system guarantees “training support for all”. A common training base promotes general acculturation while career campuses offer specific training for each profession, such as that of banking advisor.
For Stéphanie Poussou, it is important that the GEPP adopts this granular approach. AI impacts each position differently depending on the employee’s missions. Training should not be uniform for all employees but personalized and adapted to the profile and the profession. How will this profession evolve in three years? What skills will then be required? How can we close the gap with current skills through training actions?
For Marc Trilling, AI raises the question of agility and velocity and therefore the ability of companies to transform quickly. Those that eliminate jobs and replace humans with machines can certainly accelerate certain transitions but send a negative message about the value placed on human capital. On the other hand, “organizations that know how to use AI to relieve their employees of low-value-added tasks and reallocate this human time where it is really needed”, will be winners in his eyes.
