# Artificial intelligence in the workplace: a living systematic review protocol on worker safety, health, and well-being implications

## TL;DR

This paper outlines a living systematic review protocol to study how AI systems in workplaces affect worker safety, health, and well-being.

## Contribution

The study introduces a living systematic review approach to continuously assess AI's evolving impact on workers.

## Key findings

- The review will examine how AI system design and adoption influence worker safety and well-being.
- It will explore how social position affects the relationship between AI and worker outcomes.
- Findings will be used to develop strategies for responsible AI adoption in the workplace.

## Abstract

Advancements in artificial intelligence (AI) are transforming employment and working conditions in ways that shape the safety, health, and well-being of workers. We describe a protocol for a living systematic review (LSR) that will examine the interrelationship between AI systems, employment and working conditions, and worker safety, health, and well-being. Research questions are:What types of AI systems are being used within workplaces and how do their design and adoption impact worker safety, health, and well-being?How do a worker’s employment and working conditions affect the relationship between the adoption of AI systems and worker safety, health, and well-being?How does a worker’s social position (e.g., age, gender, race, disability) shape the interrelationship between AI systems at work, employment and working conditions, and their safety, health, and well-being?

What types of AI systems are being used within workplaces and how do their design and adoption impact worker safety, health, and well-being?

How do a worker’s employment and working conditions affect the relationship between the adoption of AI systems and worker safety, health, and well-being?

How does a worker’s social position (e.g., age, gender, race, disability) shape the interrelationship between AI systems at work, employment and working conditions, and their safety, health, and well-being?

A comprehensive search of primary qualitative and quantitative research will be conducted. MEDLINE, Embase (OVID), PsycINFO (OVID), and Web of Science will be searched every six to twelve months using database-specific terms and keywords. Title/abstract and full-text screening will be completed independently by two reviewers. Relevant articles will be quality appraised using a mixed method assessment tool adapted for studies of AI. Medium and high-quality studies will be synthesized using a best evidence synthesis approach. To ensure relevancy, applied workplace and AI stakeholders will provide feedback at all stages of the LSR process through dissemination excluding quality appraisal. Annually, we will evaluate the appropriateness of the review process (e.g., frequency of searches, requirement to refine research questions, utility of continuing LSR). Any amendments to protocols will be documented.

This LSR will provide timely and evolving evidence on the implications of AI in the workplace that will be disseminated through a publicly available living review dashboard. We will capture the emerging impact AI has on workers. Findings can be used to develop strategies to minimize AI’s potential workplace harms while amplifying its potential benefits, address emerging worker inequities, and inform ongoing discussions regarding responsible and safe AI adoption.

PROSPERO CRD42024625501.

The online version contains supplementary material available at 10.1186/s13643-025-03000-0.