Artificial Intelligence–Driven Contractual Conflict Management in the AEC Industry: Mapping Benefits, Practice, Readiness, and Ethical Implementation Strategies

Contracting parties in the architectural, engineering, and construction (AEC) industry often miss out on the benefits of constructive contractual conflicts and instead face adverse outcomes due to destructive ones. Studies suggest that embracing artificial intelligence (AI) can improve the management of destructive contractual conflicts. However, a gap remains in capturing a holistic view of applying an AI–driven contractual conflict management (CCM) system to manage these conflicts before they escalate into claims and disputes. Thus, this study aims to fill this research gap by answering three research questions: RQ1: What are the perceived benefits of using AI in managing contractual conflicts in the AEC industry? RQ2: What is the current state of the AEC industry regarding managing contractual conflicts and implementing an AI–driven CCM system to address these conflicts? RQ3: What core elements and strategies are needed to implement an AI–driven CCM system in the AEC industry? Data collected from 32 experts through semistructured interviews revealed that AI could aid in data analysis, conflict prediction, operational efficiency, decision analysis, negotiation, and collaboration processes within the industry. Despite facing several challenges in managing contractual conflicts, the industry is at a critical stage in implementing an AI–driven CCM system. This study identified and discussed the importance of implementing a legally framed, trustful, prevention-oriented, and human-centered AI–driven system. It also discussed the associated ethical issues and proposed frameworks to ensure a lawful, trustworthy, and human-centered AI–driven CCM system. Furthermore, it proposed an iterative implementation strategy, including formulation, context adaptation, piloting, optimization, and large-scale implementation. This study outlines theoretical and practical contributions and suggests potential research areas to address the study’s limitations.