Key points are not available for this paper at this time.
With the growing adoption of artificial intelligence (AI) in a wide range of real-world applications, eXplainable artificial intelligence (XAI) has become a rapidly evolving area of research for responsible AI development 1 , 2 . XAI techniques aim to make complex models more interpretable and trustworthy so that human users can understand how decisions are made and determine when a model’s predictions can be appropriately trusted 3 . One component of XAI is data explainability, as the composition, quality, and representativeness of training data fundamentally shape model behavior 4 . Transparency at the dataset level is particularly important in high-stakes domains, such as healthcare, where the use of biased datasets in model development has serious implications for clinical decision making 5 , 6 .
Li et al. (Wed,) studied this question.