Workplace transformation: Employer obligations in EU AI-driven restructuring scenarios, with examples from France and Germany

The rapid development of artificial intelligence (AI) and its increasing use in companies is leading to profound changes in the workplace. New AI systems are transforming job profiles or rendering existing roles obsolete. In some sectors, this may result in a surplus of labour and thus – as numerous practical examples show – lead to workforce restructuring, including staff reductions.

Before implementing such restructuring measures, companies must, in many jurisdictions, observe the information and consultation rights of employee representatives. Some of these rights apply specifically to the use of AI, while others apply more generally to restructuring processes. Depending on the country, breaches may result in, among other things, legally enforceable injunctions or even render the implemented measures (e.g. redundancies) invalid. In some jurisdictions, failure to comply with mandatory information and consultation procedures where these are required gives rise to criminal liability (e.g. for the offence of obstruction (“délit d’entrave” in France). 

Furthermore, consultation procedures with employee representatives can be time-consuming and require careful preparation. Finally, certain information obligations may arise at a very early stage, even before a company has acquired the license for a new AI system. Companies should therefore assess which information and consultation rights apply before acquiring new AI systems. 

However, discussions between employers and employee representatives regarding AI systems can also provide an opportunity to agree on the key framework for the future use of AI within a company. This could include targeted skills development initiatives and training programmes to lay the foundations for the sustainable development of employees’ AI skills.

As an example of a possible participation process, the most important participation rights of a German works council are outlined below, with some corresponding aspects of French law. Employers must observe these rights prior to the deployment of a new AI system and when implementing staff reductions triggered by AI. 

Right to information and consultation during the planning phase (Section 90(1)(3) BetrVG (German Works Constitution Act))

Employers must inform the works council well in advance of any planned deployment of AI, as well as its resulting effects on working methods and processes. They are further required to consult with the works council on the effects on employees, particularly with regard to the nature of their work and the resulting demands placed on them. This consultation must take place in good time, so that the works council’s proposals and concerns can be taken into account during the planning stage.

Similarly, in France, companies with more than 50 employees must inform and consult their works council (Comité Social et Economique, or CSE) before introducing new technologies (Article L. 2312-8 of the French Labour Code). 

French case law on the implementation of AI systems is still limited, but it generally suggests that deploying AI will require prior consultation with the CSE where it has a meaningful impact on employees. Courts tend to look at factors such as changes to working conditions (including the nature and volume of tasks), the number of employees affected, and whether the project is long-term in nature. In practice, consultation may be required whenever an AI system is likely to affect working conditions, for example by impacting employees’ health and safety, roles, skills, pay or training, or lead to staff reductions.

The consultation process must be initiated as soon as the project is sufficiently defined and before any implementation steps are taken by the company. That said, certain court rulings have waived the obligation to consult where the introduction of a new AI system was carried out solely as an optional experimental phase involving a limited number of employees.

Right of co-determination prior to the deployment of a new AI system (Section 87(1)(6) BetrVG)

When AI systems are capable of recording or evaluating the behaviour or performance of employees, the works council’s right of co-determination under Section 87(1)(6) BetrVG is triggered. German labour courts interpret this right of co-determination very broadly, with the result that virtually any use of AI falls under this provision and is subject to co-determination, regardless of whether the employer actually intends to monitor performance and behaviour of the employee. Companies may only introduce AI systems once the entire co-determination procedure has been completed. This process can take from a few weeks to several months. Ideally, a framework works agreement on the use of AI is already in place, which facilitates negotiations with the works council.

Introduction of AI as a change in operations (Section 111 BetrVG)

Whether the introduction of a new AI system constitutes a change in operations subject to co-determination and triggers further obligations for companies (see below) depends on the nature of the changes in the workplace and the consequences for employees. 

Widespread use of AI may indirectly lead to downsizing or (partial) closure of an operation. This covers all cases in which the introduction of AI results in significant job losses. A reduction in operations is assumed if the number of employees in the business is permanently and substantially reduced. This may occur, for instance, if AI were to take over repetitive tasks that were previously carried out by most employees. In extreme cases, such as when work processes are completely taken over or business models are entirely displaced, this may lead to the closure of the operation. 

However, even without staff reductions, AI can trigger a change in operations, for example if the organisation of the business changes fundamentally or key work processes are reorganised. This may be the case if decision-making powers are transferred to AI systems and employees receive their work instructions generated by AI. Finally, a change in operations subject to co-determination may also arise if one or more AI systems lead to fundamentally new working methods and production processes. The ongoing improvement of work processes is not covered by this category – instead the decisive factor is whether a qualitative leap in development occurs, rendering previous tasks entirely obsolete.

Reconciliation of Interests and Social Plan (Section 112 BetrVG)

If a change in operations is planned, the employer must provide the works council with comprehensive information and negotiate a reconciliation of interests (Interessenausgleich) as well as a social plan (Sozialplan). The reconciliation of interests determines whether, when and how the planned changes will be implemented. The social plan is intended to mitigate the economic disadvantages for employees. While traditional social plans often provide for severance pay or other compensation, AI-related restructuring offers an opportunity to focus on upskilling employees. By offering targeted training programmes, companies can maintain their employees’ employment prospects and avoid the high costs that would arise from extensive social plans or new hires. Furthermore, this enables companies to fulfil their obligations under Article 4 of the AI Act to develop the required AI literacy within the organisation.

AI as an opportunity for skills development and professional growth

Therefore, the use of AI offers companies not only technical advantages, but also the opportunity to develop employees in a targeted manner and position themselves for the future. Rather than simply replacing jobs, AI creates new roles and skill requirements. It is therefore worthwhile not to focus exclusively on rationalisation, but to invest early in training, reskilling and upskilling initiatives. 

A suitable means of achieving this can be:

• In Germany, a skills development social plan (Qualifizierungssozialplan), which prepares employees for new tasks and secures their employability by providing further training and retraining, rather than relying primarily on redundancy packages.
• In France, addressing the AI topic through workforce planning and career management frameworks (gestion des emplois et des parcours professionnels – GEPP). This allows employers and employee representatives to anticipate the impact of AI on jobs, skills and workforce structure, and to support employees through training and reskilling initiatives as roles evolve.

Companies that incorporate skills development programs as an integral part of their transformation strategy not only promote acceptance of AI but also enhance the innovative strength and future viability of the entire organisation. In this way, AI does not become a source of uncertainty, but rather the starting point for a modern, learning-oriented and resilient working environment.

Recommendation

AI is transforming the world of work rapidly and fundamentally. When introducing AI, companies should work closely with the works council to take into account its respective participation rights, ensure transparency from the outset and develop viable solutions jointly. In France, a number of companies have already addressed this topic by entering into social dialogue agreements aimed at organising the works council’s involvement in the roll-out of new AI systems. For example, by setting up an ad hoc AI committee at works council level and establishing guidelines on the information to be disclosed with the works council and the frequency of its consultations, depending on the stage of the new AI deployment project.

 Employee training should also form a central pillar of the transformation strategy. 

This will ensure that AI does not become a source of conflict, but a driver for further development and innovation.

For information on our AI practice visit Artificial Intelligence.