Construction and demolition activities produce a large and growing share of global solid waste, creating environmental, economic, and regulatory challenges for the built environment. Traditional waste management practices in construction are often reactive, fragmented, and decision-making is constrained by manual processes and limited data, which leads to inefficient material reuse and higher carbon footprints. This paper proposes an AI-supported Decision-Making Framework (AI-SDMF) designed to enable proactive, data-driven, and sustainable construction waste management. The framework combines a data collection layer (site sensors, BIM and project lifecycle data, and waste manifests) with an AI analytical engine for classification, waste-quantity forecasting, and diversion-path prediction. Outputs feed a Decision Support System (DSS) that evaluates trade-offs across cost, embodied CO₂, and circularity metrics, while a feedback and optimization module (reinforcement/adaptive learning) continuously improve recommendations. A conceptual case scenario demonstrates how AI-SDMF can reduce waste generation, increase reuse and recycling rates, and lower environmental impacts compared with conventional approaches. The framework aims to aid contractors, waste managers, and policymakers in making transparent, timely, and sustainable decisions in construction waste management.