Kaggle stroke dataset found to contain celebrity images instead of medical scans, raising concerns about medical AI data quality

Summary

A statistician discovered that a Kaggle dataset used to train clinical AI models for stroke detection contains celebrity images (Sylvester Stallone as Rambo, George Clooney, Angelina Jolie, Daniel Craig) instead of actual medical brain scans. The article highlights widespread problems with medical AI datasets, including mislabeled data, privacy violations, and low-quality training data that could lead to dangerous clinical outcomes. It examines how poor data practices in medical AI research undermine the reliability and safety of clinical prediction models for conditions like stroke, diabetes, and COVID-19.