Digital Twin AI: Opportunities and Challenges from Large Language Models to World Models
Rong Zhou, Dongping Chen, Zihan Jia, Yao Su, Yixin Liu, Yiwen Lu, Dongwei Shi, Yue Huang, Tianyang Xu, Yi Pan, Xinliang Li, Yohannes Abate, Qingyu Chen, Zhengzhong Tu, Yu Yang, Yu Zhang, Qingsong Wen, Gengchen Mai, Sunyang Fu, Jiachen Li, Xuyu Wang, Ziran Wang
2026-01-07
Summary
This paper explores how artificial intelligence is being used to make digital twins – virtual copies of real-world things – much smarter and more capable. It's about moving beyond just *showing* what's happening with something to actually *thinking* and *acting* on that information.
What's the problem?
Traditionally, digital twins were mostly used for simulations, like predicting how a machine might break down. However, they weren't very good at adapting to changing situations or making decisions on their own. The paper identifies a need for a clear way to understand how AI can be integrated throughout the entire lifespan of a digital twin, from its creation to its ongoing operation, and addresses challenges like making these AI systems understandable, reliable, and able to handle complex situations.
What's the solution?
The researchers created a four-step framework to show how AI can be used at every stage of a digital twin’s life. First, they use AI to build the initial digital model, combining traditional physics-based methods with newer AI techniques. Second, they focus on keeping the digital twin updated with real-time data from the physical object it represents. Third, they use AI to predict problems, find anomalies, and optimize performance. Finally, they explore using advanced AI like large language models to allow the digital twin to manage itself and even come up with new solutions. They looked at examples across many fields like healthcare and manufacturing to see how this framework works in practice.
Why it matters?
This work is important because it provides a roadmap for building truly intelligent digital twins. These aren't just tools for looking at data; they can become proactive problem-solvers and even autonomous systems. This has huge potential for improving efficiency, reducing costs, and creating new innovations in many different industries, but it also highlights the need to develop these systems responsibly, ensuring they are trustworthy and explainable.
Abstract
Digital twins, as precise digital representations of physical systems, have evolved from passive simulation tools into intelligent and autonomous entities through the integration of artificial intelligence technologies. This paper presents a unified four-stage framework that systematically characterizes AI integration across the digital twin lifecycle, spanning modeling, mirroring, intervention, and autonomous management. By synthesizing existing technologies and practices, we distill a unified four-stage framework that systematically characterizes how AI methodologies are embedded across the digital twin lifecycle: (1) modeling the physical twin through physics-based and physics-informed AI approaches, (2) mirroring the physical system into a digital twin with real-time synchronization, (3) intervening in the physical twin through predictive modeling, anomaly detection, and optimization strategies, and (4) achieving autonomous management through large language models, foundation models, and intelligent agents. We analyze the synergy between physics-based modeling and data-driven learning, highlighting the shift from traditional numerical solvers to physics-informed and foundation models for physical systems. Furthermore, we examine how generative AI technologies, including large language models and generative world models, transform digital twins into proactive and self-improving cognitive systems capable of reasoning, communication, and creative scenario generation. Through a cross-domain review spanning eleven application domains, including healthcare, aerospace, smart manufacturing, robotics, and smart cities, we identify common challenges related to scalability, explainability, and trustworthiness, and outline directions for responsible AI-driven digital twin systems.