Healthcare
Hospitals report patient-data incidents involving AI tools
Three academic healthcare organizations reported incidents involving patient information being entered into external or unauthorized AI tools, renewing concerns over privacy and governance.
Three academic healthcare organizations reported incidents in 2026 involving patient information being entered into external or unauthorized AI tools. One example cited in the supplied report involved Weill Cornell Medicine, which disclosed on November 5 that patient health information had been entered into AI tools without authorization during a radiology research study conducted between June 10 and June 20. The incidents highlight a governance problem created by rapid adoption of generative AI: staff or researchers may use external systems before institutional controls, training, and approval processes are fully established. The issue is significant because healthcare data can contain highly sensitive identifying and clinical information. Unauthorized uploads can create privacy, security, compliance, and trust risks even when users are trying to accelerate research or administrative work. The report frames the incidents as evidence that hospitals need clearer oversight of AI use and stronger safeguards around data handling. It does not describe a new clinical AI product, but it is directly relevant to healthcare AI adoption because it concerns how organizations manage real-world use. The cases also show that AI risk extends beyond algorithmic performance to user behavior and institutional policy. The supplied report does not establish that all incidents involved the same tools, causes, or consequences.
Original source: Becker's Hospital Review