Is AI the Solution or the Problem? Empirical Evidence on Risk Evolution in Construction Project Management

The integration of artificial intelligence (AI) into construction project management represents a transformative shift that fundamentally changes traditional risk landscapes while introducing unprecedented challenges. This study aims to (1) analyze the differences in traditional risks before and after AI implementation, and (2) examine variations in perceptions of AI-specific risks across different professional roles. A mixed-methods approach combining questionnaire surveys and expert interviews was employed. The study identifies six traditional risk categories and 14 AI-specific risks. Results demonstrate that communication risk and health and safety risk achieve statistically significant reductions following AI implementation, while confirming that traditional risk management remains indispensable in construction projects. Unauthorized AI access, AI hallucinations, and misinformation and manipulation emerge as the most critical concerns among AI-specific risks. The study also reveals that project directors and construction managers perceive misinformation and manipulation risks and AI hallucinations as more critical than project executives. This study contributes to the establishment of a comprehensive framework for assessing both traditional and AI-specific risks, highlighting how AI integration simultaneously mitigates some risks while introducing novel challenges. These findings indicate that despite AI’s effectiveness in mitigating certain risks, the persistence of traditional risks and emergence of AI-specific risks necessitate hybrid strategies that integrate AI capabilities with human decision-making and conventional risk management approaches. Furthermore, the findings provide an empirical foundation for developing targeted mitigation strategies and offer crucial guidance for industry professionals navigating the evolving risk landscape in the AI era.