- Workload, pace and interruptions
- Autonomy and discretion
- Skills, quality and meaning
- Monitoring and performance assessment
- Teams and social relations
- Occupational change and job security
CSE · AI AT WORK · OPERATIONAL FRAMEWORK
CSE and artificial intelligence at work: assess a project before issuing an opinion
A practical framework to help the Social and Economic Committee (CSE), the French employee-representation body, obtain useful information, examine effects on work and make missing safeguards visible.
10 essential questions7 review blocksNo worker scoring
WHY IT MATTERS
Why should the CSE engage with AI at work?
Direct answer
An AI deployment does not merely change a tool. It can redistribute decisions, targets, verification work, autonomy, skills, monitoring and accountability. Where consultation is required, the CSE can examine these transformations collectively while choices remain open.
OCCUPATIONAL HEALTH
Can AI at work create psychosocial risks?
Direct answer
Yes, depending on its use and organisation. AI can intensify work, reduce discretion, increase monitoring, weaken meaning at work or shift accountability. These risks must be assessed through real work.
01 / MINIMUM INFORMATION
The ten questions to ask before the opinion
For each question, indicate whether the employer’s answer is documented, partial or absent.
01Purpose
What problem are we trying to solve with AI?
A precise need, alternatives considered and the expected benefit.
02Capabilities
What can the system actually do?
Available functions, not only the announced use case.
03Decision
Which decisions can it make or influence?
Distinguish assistance, recommendation, scoring and decision.
04Autonomy
Which actions can it perform without human intervention?
Identify actions, scope and frequency.
05Access
Which data and tools can it access?
Worker data, email, files, internet and business software.
06Scope
Which workers and occupations will be affected?
Exposed groups, transformed tasks and particular situations.
07Real work
What effects are expected on work?
Workload, pace, autonomy, skills, teams, monitoring and employment.
08Appeal
How can a human challenge, correct or stop the system?
An identified person with the necessary time, information and authority.
09Incidents
How will unexpected effects be detected?
Reporting channel, logging, accountable owner and feedback.
10Suspension
Under which circumstances will the system be suspended?
Written stopping thresholds and a tested rollback procedure.
02 / POWER GIVEN TO THE SYSTEM
Qualify the system’s real autonomy
The higher the level, the stronger the justification, control, monitoring and reversibility must be.
Who can stop the AI? Does that person genuinely have the time, information, skills and authority required to do so?
Could the project alter organisation, workload, autonomy, pace or working conditions?
03 / COMPLETE FRAMEWORK
Seven blocks, in decision order.
Start with the system, examine real work and finish with a reasoned conclusion.
Understand the system
Model, provider, version, hosting, memory, tools and agents.
See transformed work
Workload, autonomy, skills, relations, assessment and employment.
Qualify autonomy
From information to organisational delegation.
Prove human control
Time, information, skill, authority and genuine stopping power.
Prepare for incidents
Reporting, severity, analysis, suspension and rollback.
Organise reassessment
New model, data, tools or affected population.
State a conclusion
What is documented, missing and must be strengthened.
YOUR SUMMARY
Additional information required
Some answers are still missing to understand the system and its effects.
04 / ESSENTIAL ANSWERS
Frequently asked questions about the CSE and AI.
Direct answers on consultation, occupational health and the information to obtain.
Must the CSE be consulted before every use of AI?
No. An individual’s occasional use does not by itself trigger consultation. An employer-led introduction may, however, fall within prior information and consultation requirements depending on company size, capabilities and effects on work.
Can the CSE seek input from the occupational physician?
Yes, particularly where workload, autonomy, pace, organisation or psychosocial risks may be affected.
What information should the employer provide?
Purpose, actual capabilities, influenced decisions, autonomy, accessible data and tools, affected workers, expected effects, human control, incident monitoring and suspension criteria.
05 / LEGAL FRAMEWORK
A legal preparation tool
Not every use of AI automatically triggers consultation. Company size, the nature and capabilities of the project, and its effects must be assessed case by case under the French Labour Code.
The CSE should obtain information before the decision where the project falls within its information-and-consultation remit. This tool supports preparation; it does not replace legal advice tailored to the situation.