Haggenmüller et al. analyzes whether or not having a “privacy-preserving federated learning approach” is a viable diagnostic option for AI-based melanoma detection (Haggenmuller and Brinker, 2024). This approach aims to protect the sanctity and security of patient data while also allowing for more accurate diagnosis. This study was completed using 6 German university hospitals between April 2021 and February 2023, where a test dataset was employed to analyze a single-arm diagnostic study for differentiation between melanoma and nevi. The researchers inspected these whole-slide images via an AI classifier, without interference to routine clinical care efforts. The federated approach yielded significantly worse performance in comparison to the classical centralized approach on a holdout data set according to “area under the receiver operating characteristic curve” (AUROC) score. Nevertheless, the federated approach performed better than the classical centralized approach when looking at an external dataset. These findings suggest a federated learning approach may be useful in classifying the severity of pigmented lesions (i.e., making distinctions between melanomas and nevi).