The integration of artificial intelligence (AI) into construction safety management presents considerable potential for improving safety performance. However, it remains unclear how the differing cost–benefit expectations, risk perceptions, and strategic choices of relevant stakeholders influence the willingness of key players, i.e., government regulatory authorities (GRA), AI technology suppliers (AITS), and construction contractors, to invest in AI-based safety solutions. To address this gap, this study develops an evolutionary game model that incorporates prospect theory and mental accounting to analyze the AI investment willingness and strategic interactions of key players. The findings first reveal that although stakeholders differ in their initial investment willingness, they ultimately converge to a stable equilibrium: GRA maintain a leadership role focused on public welfare, AITS adapt quickly to profitable innovation, and contractors advance cautiously until returns become evident. Secondly, variations in profitability, costs, and incentive allocation significantly influence how quickly and in what direction stakeholders adjust their investment strategies, emphasizing the need for mechanisms that account for each stakeholder’s sensitivity. Lastly, the paper found that adaptive regulatory frameworks and equitable incentives reduce opportunism, foster collaboration, and support long-term AI adoption in construction safety. This study enhances understanding of stakeholder decision-making in AI investment and provides practical insights for designing effective incentives and regulatory strategies.
Tian et al. (Mon,) studied this question.