On January 2, 2026, a research team led by Professor Bing Sheng from the School of Computer Science and Engineering at Shanghai Jiao Tong University, in collaboration with Professor Tianyin Huang from the School of Medicine at Tsinghua University and other interdisciplinary experts, published a viewpoint article in The Lancet Digital Health (IF=24.1), one of the world’s leading medical journals. The study, titled “Can generative artificial intelligence empower target trial emulations?” (DOI: 10.1016/j.landig.2025.100950), presents a novel framework that integrates generative artificial intelligence (AI) with digital twin technology to revolutionize target trial emulation (TTE)—a method used to simulate randomized controlled trials (RCTs) using real-world observational data. The research addresses a long-standing challenge in clinical research: the limitations of traditional RCTs, which, despite being the gold standard for establishing causal relationships between medical interventions and patient outcomes, are often prohibitively expensive, time-consuming, and restricted by narrow inclusion and exclusion criteria. These constraints limit generalizability and make RCTs ethically or practically unfeasible for certain interventions. TTE offers a promising alternative by applying RCT principles to existing observational datasets—such as electronic health records—thereby enabling more transparent, scalable, and inclusive evaluations of real-world effectiveness. However, TTE faces three major barriers: complex study design requiring high expertise; persistent confounding factors that prevent accurate estimation of unobserved counterfactual scenarios (e.g., what would happen if a patient received a different treatment); and insufficient data for rare diseases or vulnerable populations, leading to low statistical power. To overcome these challenges, the team proposes a three-pronged approach powered by generative AI: Intelligent Study Design: Generative AI can automatically translate clinical concepts into standardized trial definitions, extract relevant variables from electronic health records, and generate structured, compliant research protocols—significantly lowering the technical barrier for researchers and democratizing access to TTE. Counterfactual Simulation: By leveraging digital twin and medical world model concepts, generative AI can simulate unobservable clinical scenarios—such as alternative treatment regimens, dose adjustments, or treatment timing—thereby enhancing causal inference and improving the validity of real-world evidence. Synthetic Data Generation: The technology can create high-fidelity synthetic datasets that mirror real-world data distributions, effectively expanding sample sizes for underrepresented groups and rare conditions. This improves statistical power and ensures greater representativeness, advancing equitable and inclusive medical research. The authors also emphasize the need for caution. Risks include potential privacy breaches in synthetic data, amplification of biases present in original datasets, and reduced model interpretability due to the “black box” nature of AI. To mitigate these, the team calls for clear technical standards, transparency protocols, and robust accountability mechanisms. They advocate for cross-disciplinary collaboration among clinicians, trial designers, data scientists, and regulators to ensure ethical, reliable, and reproducible implementation. This work builds on the team’s earlier publication in Nature Biomedical Engineering, which highlighted the dual potential of synthetic data—its ability to overcome data privacy, annotation cost, and data silo challenges, while also posing risks of introducing bias or “data toxicity.” The current study refines this vision by proposing a “real-world data as foundation, synthetic data as supplement” integration strategy, offering a practical roadmap for translating AI innovations into clinical practice. The research was conducted by Sheng Bing’s team at Shanghai Jiao Tong University and Huang Tianyin’s team at Tsinghua University, with co-authors including Zhiyu Guan (Shanghai Jiao Tong University), Dian Zeng (Tsinghua University), and Hua Ting Li (Shanghai Jiao Tong University). It was supported by multiple national and regional grants, including the National Natural Science Foundation of China, the National Key R&D Program, the National Major Project on Chronic Diseases, and the Shanghai Science and Technology Major Project. The team has long focused on AI and medicine integration, particularly in intelligent diagnosis for metabolic and eye diseases. This study not only provides a new tool for generating medical evidence in scenarios where RCTs are impractical but also advances the shift from data-driven to knowledge-driven clinical research. It strengthens China’s technological competitiveness in healthcare, supports the development of a self-reliant medical data ecosystem, and contributes to the national “Healthy China” strategy. Looking ahead, the team plans to deepen interdisciplinary collaboration with clinical institutions and regulatory bodies to standardize and scale the application of generative AI in TTE, ensuring that cutting-edge technology translates into tangible benefits for patients and drives sustainable innovation in medical science. Corresponding Author Profile: Bing Sheng, Ph.D., is a professor and doctoral supervisor at the Institute of Computer Applications and the Key Laboratory of Artificial Intelligence, Ministry of Education, School of Computer Science and Engineering, Shanghai Jiao Tong University. He leads the National Natural Science Foundation of China Young Scientist Fund (Category A) project and serves as Deputy Director of the Shanghai International Joint Laboratory for Smart Prevention and Control of Metabolic Diseases under the Belt and Road Initiative. He is a recipient of China’s National High-Level Youth Talent Program. His research focuses on virtual reality and intelligent medicine. He has served as Chair of the ISBI 2020 and MICCAI 2022–2024 International Fundus Image AI Reading Competitions and as Managing Editor of The Visual Computer. In the past five years, he has published 69 papers as first or corresponding author in top journals including JAMA, Nature Medicine, The Lancet Digital Health, IEEE TPAMI, and IJCV. He has received numerous awards, including the Shanghai Science and Technology Progress Special Prize, two SAIL Awards at the World Artificial Intelligence Conference, the Outstanding Contribution Award from the Computer Graphics Society, the Top Ten Innovations in Medical Engineering in China, the 2024 China’s Most Important Medical Advances (awarded by the Chinese Academy of Medical Sciences), and the Top Ten Original Discoveries in Ophthalmology in China. He currently leads key national projects, including a sub-project under the National Key R&D Program, the National Natural Science Foundation, and projects funded by the Shanghai Science and Technology Commission.