Professor Huang Tianyin, the dean and chair professor of the School of Medicine at Tsinghua University, alongside a team of experts that includes Associate Professor Wang Yaxing from the Eye Center and the Beijing Advanced Innovation Center for Biomedical Engineering, has jointly published a groundbreaking study on the generalization capabilities of basic AI models using Asian datasets. The research highlights the constraints faced by these models in terms of data diversity, underscoring the necessity to diversify data sources and foster global research partnerships to overcome these hurdles. By evaluating the performance of the RETFound basic model against the conventional Vision Transformer model on Asian data, the study found no statistically significant difference in their efficacy in diagnosing glaucoma, coronary heart disease, and predicting stroke risk. This investigation underscores the critical role of global collaborative research in enhancing the universality and fairness of medical basic models.