Applicability of artificial intelligence models for selection of ADR methods to the settlement of contractor-related variation disputes in building construction projects in Sri Lanka

Abstract

Construction disputes in Sri Lanka are highly technical in nature and differ significantly from general commercial disputes, necessitating fast and cost-effective resolution methods. Due to the drawbacks of litigation such as high cost, delays, and complexity, Alternative Dispute Resolution (ADR) methods have become increasingly important, offering advantages like speed, affordability, fairness, simplicity, flexibility, confidentiality, and minimal delays. In building construction projects, variations are common and often lead to disputes, with contractor-related variation disputes being particularly prevalent and impactful. Effective resolution of these disputes is essential for successful project completion. This research aims to propose a guideline that utilizes Artificial Intelligence (AI), including machine learning and deep learning techniques, to select the most suitable ADR method for settling contractor-related variation disputes in building construction projects in Sri Lanka. The study adopted a qualitative approach, conducting expert interviews with ten professionals selected through purposive sampling, all of whom had experience in ADR and construction variations. Data were analysed using code-based content analysis to identify appropriate ADR methods. The research findings led to the development of a guideline integrating AI tools to support decision-making in ADR method selection. This guideline provides valuable insights for industry practitioners, enabling more efficient and effective resolution of contractor-related variation disputes within Sri Lanka’s building construction sector.