🔍 Did you imagine that AI could be deployed in your company without the Works Council (CSE) having a say? Think again; that would be underestimating our beloved France! 🛑
This is the essence of the interim order (RG No. 24/01457) issued on February 14, 2025, by the Nanterre Judicial Court, marking a turning point in the legal framework for AI deployment in companies.
🔍 Facts and Decision
A company was forced to suspend the deployment of AI applications for bypassing the mandatory consultation of the CSE. The employer’s argument that the tools were merely in an “experimental phase” did not convince the judge, who considered that this phase already involved the use of new software by all affected employees, thus constituting a “manifestly unlawful disturbance.”
📊 Key Points:
Immediate suspension of AI application deployment until the completion of CSE consultation.
Insufficient consultation, with the CSE deeming additional documents necessary.
Premature deployment, as the company introduced AI tools before the CSE had rendered its opinion.
Obsolete pilot phase, with the court considering that even a pilot phase involved partial use by employees.
Fine of €5,000.
🤖 Towards the Emergence of an Internal AI Regulation Ecosystem ❓
In a context where the external regulatory framework (European AI Act, CNIL’s 2025-2028 Strategic Plan) is gradually strengthening, the entire internal ecosystem of the company is called upon to mobilize.
We can envision the emergence of transversal “AI ethics committees” bringing together legal experts, IT departments, Data Protection Officers, and CSE representatives to validate use cases before any deployment.
⚖️ Legal departments will therefore need to develop specific expertise in “AI compliance” to:
Classify AI systems according to their risk level.
Develop standardized impact assessment processes.
Formalize a register of algorithmic processing.
Create technical and social documentation suitable for CSE consultations.
The Nanterre case law, overturning decisions of the Criminal Chamber of the Court of Cassation, which considered that the introduction of new software did not constitute a significant project requiring CSE consultation (Cass. crim., May 3, 1994, No. 93-80.911; Cass. crim., September 13, 2005, No. 04-86.887), likely inaugurates a new era where AI regulation in companies will rely on shared governance among the CSE (guardian of working conditions), the legal department (guardian of compliance), the IT department (guardian of security), and business units (guardians of utility).
💬 What about your company? Has your CSE been duly consulted before deploying an AI tool 😉?
Source : https://foc.media.fo-cadres.fr/1739902905394_7d61d0dfc3.pdf
