Expert-informed causal mapping of generative AI adoption for net-zero built environment: a paradox theory perspective

The rapid diffusion of Generative Artificial Intelligence (GenAI) is increasingly framed as a key enabler of net-zero transitions in the built environment. Yet, its adoption simultaneously introduces persistent and interdependent tensions that complicate decarbonization, governance and project delivery. This study aims to answer the three important questions: what paradoxes emerge, why they exist and how they interact. Using the DEMATEL method based on the perspectives of 15 experts in Vietnam, this study develops a causal mapping of these paradoxes grounded in paradox theory. The DEMATEL results reveal three key findings. Firstly, the automation–human capability paradox is the most prominent systemic tension, reflecting how GenAI-driven decarbonization can undermine essential human expertise. Secondly, institutional and governance paradoxes operate as upstream drivers, indicating that many GenAI-related challenges stem from governance misalignment rather than technological limitations. Thirdly, the digital energy paradox emerges as a downstream effect, showing that rising computational energy demand is largely a consequence of unresolved governance, standardization and data-management tensions. This study contributes to theory by introducing a causal, system-level explanation of GenAI-enabled net-zero transitions. It supports policymakers and industry stakeholders in deploying GenAI as a credible enabler of net-zero outcomes in the Vietnamese built environment and other similar countries.