AI-enabled Transport Digital Twins (TDTs) are increasingly being used for real-time traffic monitoring, prediction, infrastructure management, and intelligent decision-making. However, their continuous connectivity with sensors, vehicles, communication networks, and cloud-edge platforms creates significant cybersecurity and operational risks. Conventional security approaches generally rely on static protection mechanisms and periodic assessment, which may be insufficient for dynamic transportation environments. This paper proposes a Risk-Aware Continuous Assurance and Cybersecurity Framework (RCAC-TDT) for AI-enabled Transport Digital Twins. The proposed framework continuously monitors data integrity, Digital Twin synchronization, AI model behavior, network security, and operational conditions. A risk-aware mechanism classifies detected events according to their potential impact and dynamically determines appropriate responses, ranging from continuous monitoring to human intervention and temporary AI restriction. The framework also incorporates AI model monitoring, anomaly detection, cybersecurity controls, explain ability, and auditability. The proposed approach provides a lifecycle-oriented mechanism for maintaining the security, reliability, and operational assurance of AI-enabled Transport Digital Twins. The framework provides a foundation for future experimental validation using transportation simulation environments and real-world datasets.