Construction and demolition waste (CDW), which is increasing rapidly worldwide and causing significant environmental impacts, originates not only from building construction and demolition but also from infrastructure development, renovation projects, and roadworks. This diversity requires differentiated management strategies based on waste types and activities. CDW accounts for a significant share of global material use, waste generation, energy consumption, and emissions, and demolition waste poses particular risks to sustainable practices. Although much of this waste is recyclable, recycling rates vary widely across countries. Due to its high volume and potential for recovery, CDW is recognised as a priority waste stream in sustainable construction policies. This study investigates integrating artificial intelligence (AI) into CDW management to promote sustainable, circular construction practices. A systematic review of literature published between 2010 and 2025 was conducted using the Web of Science and Scopus databases, evaluating the development of AI-based applications over time. Results show a notable increase in publications after 2020, with CNN and Transformer architectures leading in image-based classification and segmentation, while traditional algorithms such as ANN, SVM, and Random Forest remain prevalent in prediction tasks. Key challenges include a lack of data, material heterogeneity, and limited policy integration. Future research directions focus on multimodal sensing, hybrid and explainable AI, and transfer learning. Overall, the study highlights how AI enhances decision-making, resource efficiency, and innovative construction practices, while also identifying the limitations and research gaps that need addressing to advance sustainable CDW management.