AI in French Employment Law
AI and French Labour Law
The deployment of artificial intelligence in the workplace raises novel legal questions under French labour law, intersecting with the Code du travail, data protection regulation, and the emerging EU AI Act framework. French law has traditionally afforded robust protections to employees against technological surveillance and arbitrary management decisions, and these protections increasingly apply to algorithmic management and AI-assisted human resources practices.
Worker Surveillance and the Code du Travail
French labour law imposes strict limits on employer surveillance of workers. Article L1222-1 of the Code du travail requires that employment contracts be performed in good faith. Article L1222-4 provides that no information concerning a worker may be collected by any means without the worker’s prior knowledge and without the worker having been informed of the purpose of the data collection. The Cour de cassation has consistently held that covert surveillance is unlawful unless the employer has specific grounds to suspect misconduct (arrêt Nikon, Cass. soc., 20 February 2001). An employer cannot rely on evidence obtained through hidden cameras, undisclosed geolocation tracking, or undisclosed monitoring of computer usage.
These principles extend to AI-powered surveillance tools. An employer deploying facial recognition for access control, keystroke monitoring software, or productivity tracking algorithms must inform employees and their representatives, conduct a data protection impact assessment, and ensure proportionality. The Commission nationale de l’informatique et des libertés (Cnil) has issued guidance on workplace AI, emphasising that automated monitoring must respect human dignity, not be permanent or generalised, and must not assess employee performance solely through algorithmic means without human oversight.
Algorithmic Management
Algorithmic management — the use of algorithms to assign tasks, set schedules, evaluate performance, and make or recommend disciplinary decisions — is subject to the information and consultation rights of the Comité Social et Économique (CSE). Under Articles L2312-8 and L2312-38 of the Code du travail, the employer must consult the CSE before introducing any automated data-processing system that may affect working conditions or the collection of personal data. The CSE may have recourse to an expert (expertise) at the employer’s expense to assess the implications of the proposed technology.
The EU AI Act, adopted in 2024, classifies AI systems used for employment, worker management, and access to self-employment as high-risk, requiring conformity assessment, risk management, transparency, and human oversight. French law reinforces these requirements through the Code du travail provisions on health and safety, which impose on the employer a general obligation to protect the physical and mental health of workers. Algorithmic systems that intensify work pace, reduce autonomy, or cause psychosocial risks may breach this obligation.
Profiling and Discrimination in Recruitment
AI-assisted recruitment tools — including résumé screeners, video interview analysers, and psychometric profiling algorithms — raise significant discrimination risks under French anti-discrimination law. The loi Pleven (1972) prohibits discrimination on grounds of origin, sex, family situation, age, health, disability, political opinions, trade union activities, or membership in a particular ethnic group. The loi du 27 mai 2008 transposed EU equality directives and extended protection to additional grounds. The Code pénal criminalises discrimination in employment.
AI recruitment tools may inadvertently discriminate by learning biases from historical hiring data or by using proxies for protected characteristics. French law places the burden of proof on the employer in discrimination claims: once the employee presents evidence of disparate treatment, the employer must demonstrate that its decision was justified by objective, non-discriminatory factors. The Défenseur des droits has published guidance on algorithmic non-discrimination, recommending that employers audit their AI tools for bias, ensure transparency in algorithmic decision-making, and maintain human intervention in all significant recruitment decisions.
Cnil Guidance and Enforcement
The Cnil has been proactive in addressing AI and employment questions. Its 2023 action plan on AI includes a focus on workplace uses, and it has published practical recommendations on the lawful deployment of AI in human resources. The Cnil emphasises the principles of data minimisation — employers should not collect more data than necessary — and purpose limitation — data collected for one purpose, such as access control, should not be reused for performance evaluation. The Cnil has also addressed the use of emotion recognition AI, considering it inherently problematic for employment purposes due to its lack of scientific reliability and its intrusiveness.
Enforcement powers include administrative fines of up to 4% of global annual turnover under the GDPR, injunctions to cease processing, and public warnings. The Cnil has imposed significant fines for workplace surveillance violations, including against Amazon France Logistique for its warehouse employee monitoring systems, which were found to be excessively intrusive.
Works Council Consultation
Beyond the CSE consultation requirements, French law requires information and consultation on any project likely to affect the organisation, management, or working conditions of the enterprise. The introduction of AI systems that automate task allocation, performance assessment, or disciplinary processes triggers these obligations. The CSE must receive sufficient information in good time to enable it to express an informed opinion. Non-compliance with consultation obligations may result in the suspension of the project and claims for damages.