[Relationship between certain uses of artificial intelligence and psychosocial risk factors in European work environments].

Abstract

Introduction: To examine the relationship between the use of Artificial Intelligence (AI) to assess and monitor job performance and exposure to psychosocial risk factors, as well as associated adverse health effects in the European work environment.

Method: Cross-sectional study using microdata from the 2022 "Occupational Safety and Health in Post-Pandemic Workplaces (Flash Eurobarometer)" survey (EU-OSHA) with 27252 participants. After selecting 12 dichotomous dependent variables (psychosocial risks and adverse health effects) and the presence of AI and its various uses to supervise and evaluate workers performance as independent variables, we calculated the crude and adjusted (aOR) odds ratios by sociodemographic covariates and their corresponding 95% confidence intervals (95%CI).

Results: When AI is used to monitor or control individual performance, it increases time pressure and work overload (ORa=1.5;95%CI:1.3-1.7), reduces autonomy or influence over work processes (ORa=2.2;95%CI:2.1-2.3), and erodes communication or cooperation within the organization (ORa=1.5;95%CI:1.4-1.6). It also increases the probability of reporting stress, depression or anxiety (ORa=1.5; 95%CI:1.4-1.5) and accidents or injuries (ORa=1.7; 95%CI:1.6-1.8).

Conclusions: AI as a "digital supervisor" increases exposure to psychosocial risk factors and the likelihood of health damage. This highlights the importance of considering worker well-being along with economic efficiency when implementing AI in work organizations. These results can guide labor policies to balance process optimization with healthy work environments through social dialogue.