Retraction Watch's weekend roundup highlights a notable finding: private information appears in 'most' arXiv preprints, raising significant questions about AI training data, privacy, and preprint governance. The item also covers 28 misconduct cases in China and the George Church authorship story. The arXiv privacy finding is directly relevant to P2P Review's AI coverage — preprints are a major source of AI training data, and the revelation that they contain private information adds a new dimension to the AI/research integrity debate. Retraction Watch database now exceeds 65,000 retractions, a milestone worth noting.

Source: Retraction Watch