Publications / JOEM editorial
Publication · JOEM · Occupational health and AI safety
Occupational Health Is an AI Safety Issue
An editorial co-authored with Pouran D. Faghri, published ahead of print on 10 November 2026 in the Journal of Occupational and Environmental Medicine (JOEM). It argues that occupational health is part of AI safety and proposes a framework of six requirements.
01 / The reference
A co-authored editorial published in JOEM.
What you need to find and cite it.
| Item | Detail |
|---|---|
| Title | Occupational Health Is an AI Safety Issue |
| Authors | Charles Broutin and Pouran D. Faghri |
| Type | Editorial |
| Journal | Journal of Occupational and Environmental Medicine (JOEM) |
| Published | 10 November 2026, online ahead of print (Publish Ahead of Print) |
| Language | English |
| Identifier | DOI 10.1097/JOM.0000000000003935 |
| Access | Editorial text on the journal’s website ↗ |
Broutin C, Faghri PD. Occupational Health Is an AI Safety Issue. Journal of Occupational and Environmental Medicine. Published ahead of print, 10 November 2026. doi:10.1097/JOM.0000000000003935
02 / The thesis
Occupational health is part of AI safety.
As this site puts it.
By connecting work as it is actually done, organisation and health, occupational health already has analytical frameworks and preventive measures to detect AI-related changes in work early.
Its role should not begin when workers become unwell: it should begin when tools are chosen and work is redesigned.
03 / The proposed framework
Six requirements for deploying AI without degrading work.
The framework proposed in the editorial, in the order this site presents it.
- Assess risks before deployment.
- Give workers an effective voice.
- Recognise hidden work.
- Ensure meaningful human control.
- Preserve skills and social support.
- Share benefits and monitor consequences.
This framework is proposed in the editorial; its own effectiveness remains to be evaluated. It is neither a validated method nor a regulatory standard.
04 / Into practice
For each requirement, a resource on this site.
Starting points for working on each requirement with the people concerned.
| Requirement | Resource on this site |
|---|---|
| Assess risks before deployment | Assessment grid for a workplace AI project, to complete before the pilot. |
| Give workers an effective voice | French CSE and AI: ten questions before the opinion. |
| Recognise hidden work | AI and psychosocial risks: checking AI outputs, which is rarely counted. |
| Ensure meaningful human control | Human in the loop: what supervising agents does to the supervisor. |
| Preserve skills and social support | Recommendations by audience: practice without assistance, training during working hours. |
| Share benefits and monitor consequences | Follow-up schedule and AI and jobs. |
05 / Key answers
Three questions about the editorial.
What this page tells you before you open it.
Where can the editorial be read?
On the website of the journal, the Journal of Occupational and Environmental Medicine, where it was published ahead of print on 10 November 2026. Open the editorial ↗
What does the editorial propose?
A framework of six requirements: assess risks before deployment, give workers an effective voice, recognise hidden work, ensure meaningful human control, preserve skills and social support, and share benefits and monitor consequences.
Is this framework a validated method?
No. Its own effectiveness remains to be evaluated. It is neither a validated method nor a regulatory standard.
Update log
- First version of this page. It gives the editorial’s reference and the framework as presented on the site’s homepage.