Springer Nature has retracted five publications on autism since November 16 and plans to retract 33 more. These 38 papers, conference proceedings, and book chapters relied on a problematic dataset from Kaggle. The file contains images of children's faces taken from autism websites, raising concerns about consent and data reliability. The dataset is marked as unsuitable for research because it does not meet the standards for a large controlled sample with a clinically confirmed diagnosis of autism. Publisher Tim Kersjes, head of research integrity at Springer Nature, confirmed this. A Kaggle user raised ethical concerns, but the creator of the dataset responded that the images were publicly available and served to support early detection of autism.