Healthcare
UC San Diego Health and Anthropic develop oversight for patient-facing AI
UC San Diego Health, Harvard Medical School, and Anthropic are developing an open-source system to monitor safety in patient-facing healthcare AI.
UC San Diego Health is working with Harvard Medical School and Anthropic on an open-source oversight system for patient-facing AI. The system is designed to check the safety of medical recommendations generated for patients, including advice related to medication-dose changes. The project addresses the period after an AI tool has gone live, when ongoing monitoring is needed to identify unsafe recommendations. The development matters because patient-facing systems can influence behavior outside direct clinician supervision. An open-source oversight approach could provide healthcare organizations with reusable methods for evaluating and monitoring AI-generated clinical guidance.
Original source: Becker’s Hospital Review