Identifying suitable artificial intelligence technologies for safety management in high-rise projects in Vietnam using parametric modeling and multivariate regression analysis

Abstract

Rapid urbanization in Vietnam is fueling a boom in high-rise construction projects, leading to increasingly complex occupational safety challenges. Although Artificial Intelligence (AI) solutions such as Computer Vision, Predictive Analytics, and Robotics/Drones have been successfully applied in some developed countries, widespread implementation in Vietnam still faces multiple barriers (financial, safety culture, data infrastructure, etc.). This study develops a parametric model with budget, complexity, safety culture, and institutional support as input variables to forecast the potential benefits (or “benefit scores”) of each AI technology for managing safety in high-rise construction sites. Data were collected from 12 ongoing high-rise construction projects, yielding 170 valid responses. The research employs Multivariate Analysis of Variance (MANOVA) to assess the combined impacts of these parameters on three indexes of the groups of AI technologies. The results show that budget, safety culture, and institutional support significantly explain variations in AI-driven safety benefits, whereas the project complexity index does not exhibit a noteworthy effect under the multivariate test. Separate OLS regressions for each AI score further reinforce the finding that each AI technology displays different sensitivity levels to these parameters. The study contributes theoretically by highlighting the crucial roles of cultural and managerial support factors, and practically by offering a data analysis model and guidelines for safe AI adoption in Vietnam.