From reactive to predictive prevention: how data and AI are transforming occupational safety and health
The growing availability of data and advances in machine learning techniques are expanding the possibilities for occupational safety and health (OSH), enabling a shift from a predominantly reactive approach towards more anticipatory prevention. This article examines the role of predictive models in this transformation, explaining how they work and their main applications for anticipating accidents, estimating their severity, identifying periods of increased risk, and predicting other occupational health outcomes. Particular attention is paid to the transition from prediction to preventive action, emphasizing that the value of these models lies not only in their predictive capacity but also in their use for setting priorities and supporting decision-making. The article also discusses their main limitations, including data quality and representativeness, bias, explainability, and privacy. It concludes that predictive prevention should be understood as a tool that complements professional knowledge and judgement, aimed at making better use of available information to act before harm occurs.
ODS

Xavier BarazaAssociate professor of Occupational Health and Safety and dean of the Faculty of Economics and Business at the Universitat Oberta de Catalunya. He is a researcher in the interdisciplinary research group i2TIC-IA Lab (UOC-DIGIT) and focuses his research on analysing major accidents across sectors, the evolution of occupational health and safety professionals’ skills, and modelling industrial chemical accidents. He holds a degree in chemical engineering (UB, 1999), a PhD in chemical engineering (UB, 2002) and master degrees in Experimental Chemistry (UB, 2000), Occupational Risk Prevention (UPC, 2003) and Business Administration (EEN, 2007).
Joan Torrent-SellensFull professor of Economics, director of the Digital Transformation and Governance Research Centre (UOC-DIGIT) and co-principal investigator of the Interdisciplinary Research Group in ICT and AI Lab (i2TIC-IA Lab). His research focuses on the political economy of artificial intelligence and its economic, business, labour, well-being and health, and sustainability effects on digital transformation and transition processes. He holds a degree in Economics and Business Administration (UAB, 1994), a master degree in Economic Analysis (UAB, 2001) and a PhD in Economics (UOC, 2002).
Josep-Anton Mir-TutusausAssistant professor of Big Data and Business Intelligence in the Faculty of Economics and Business at the Universitat Oberta de Catalunya. He is a researcher with the TURBA Lab group (UOC-TRÀNSIC) and focuses his research on the intersection of environmental sciences, social sciences and artificial intelligence through quantitative analysis and machine learning. His work seeks to dismantle the “immaterial” discourse of technology by analysing the physical costs (water footprint, carbon footprint and material extraction), as well as the social and labour costs of the transformation driven by artificial intelligence. He holds a PhD in Environmental Science and Technology (ICTA-UAB, 2017), a degree in biotechnology (UAB, 2012) and a master degree in Bioinformatics and Biostatistics (UOC-UB, 2023).

