Digital Twins and Agentic AI: Applications, Risks, and Governance Frameworks

The rapid evolution of artificial intelligence (AI) has transformed digital twins (DTs) from passive simulation models into systems that can support perception, reasoning, planning, and coordination, as well as action, in some settings. However, the integration of agentic AI with DTs creates governance challenges involving autonomy, cybersecurity, accountability, privacy, and goal misalignment. This paper presents a structured review of agentic-AI-enabled DTs, focusing on applications, agent architectures, risks, autonomy, human control, and governance. It uses explicit research questions and a structured classification framework to select the reviewed literature according to the application domain, agent architecture, autonomy level, risk category, level of human control, and governance mechanism. The reviewed studies indicate that autonomy is context-dependent, where higher-consequence and less-reversible decisions require stronger authorization, validation, monitoring, and intervention mechanisms. Based on the agentic-DT context, the paper presents a fusion of governance frameworks and DT applications linking governance actors, autonomy constraints, assurance mechanisms, human intervention, and recovery across the DT-agent lifecycle.