Artificial intelligence and occupational health.

Sourced guidance, guides and tools to assess an AI project before deployment, prevent its effects on health and use it with judgement in occupational health.

You are

  • An occupational physician or nurse

    AI concerns both the workers you support and your own practice. In companies, it can change workload, autonomy or the skills people draw on: changes to take into account in workers’ health monitoring and in prevention advice. In your service, it also opens up possibilities for literature searches, preparing materials or reviewing documents. This site helps you understand both dimensions, identify relevant uses and know their limits, without losing sight of data confidentiality or your professional judgement.

    Recommendations
  • An employer, manager or HR professional

    Bringing AI into a team is an organisational decision, not just a choice of tool. Who will do the remaining tasks? How much time will need to go into checking? How will objectives, skills and responsibilities change? The expected gains deserve to be examined alongside their consequences for day-to-day work. This site offers guidance and an assessment tool to prepare a project with the people concerned, anticipate difficulties and monitor the effects after deployment. The aim: to decide on the basis of what AI actually brings to your organisation, not only on its advertised performance.

    Recommendations
  • A works council member

    The effects of an AI project on employees must be open to discussion before the choices are set in stone. Task allocation, monitoring, how jobs evolve, autonomy or workload: behind a technical presentation lie questions that directly concern employee representation. This site helps you understand the intended uses, request the information you need and examine the safeguards on offer. In particular, you will find ten questions to prepare the review of a project, feed the dialogue with management and organise follow-up of the commitments made, without needing to understand the technical workings of the models.

    Recommendations
  • A prevention or ergonomics specialist

    Automating a task does not necessarily remove the workload. Activity can shift towards checking, correcting errors or handling more complex situations. Your analysis can make these shifts visible and show what they change for individuals and teams. This site examines the effects of AI through the psychosocial risk factors of the Gollac report, along with practical fact sheets to spot warning signs and define prevention measures. Resources to enrich field observation and discussion about work, not to replace them.

    Recommendations
  • An employee

    AI can affect you even if you do not use it yourself. It may play a part in how your tasks are allocated, how your work is assessed or the decisions you are asked to apply. It can also help you with some tasks, while changing what is expected of you. This site gives you pointers to understand these changes, identify what makes your work easier or harder, and put words to the difficulties you encounter. Material for discussions with your team, your manager, your staff representatives or your occupational health service, and for taking part in the choices that concern your work.

    Recommendations
  • A researcher

    An AI system’s performance is not enough to understand its effects on work and health. How can its usefulness in professional practice be assessed? Which organisational changes might explain favourable or unfavourable effects? How can a technical promise be told apart from an observed benefit? This site brings together publications and analyses to explore these questions, making methods and their limits explicit. It also addresses a complementary relationship: how the working conditions of the teams that develop and evaluate models could influence AI safety. A meeting point between occupational health, the human sciences and artificial intelligence research.

    Recommendations

Tools

  • Assess an AI project

    Nine work axes examined before the pilot, then compared with the effects observed after deployment.

    18 questions · about 10 minutes · nothing is sent

  • Works council checklist

    The questions to ask the employer before giving an opinion on an AI project, with an automatic summary.

    10 questions · summary of missing information

  • Pilot file

    The documents that make a pilot open to discussion, traceable and reversible before an AI is tested.

    8 documents · checklist

  • Prevention fact sheets

    Sheets on psychosocial risks and AI built on the six Gollac dimensions: warning signs and prevention measures.

    7 PDF sheets to download

Latest reading

All reading