A Call for Disclosure When Using AI for Patient Communications

AbstractIn the spring of 2023, the UC San Diego Health system began generating draft replies to patient messages with next-generation artificial intelligence (AI) tools as an early adopter. The rollout presented our AI governance committee with novel challenges, particularly regarding transparency in the use of AI with our patients, which we chose to make explicit with an automated disclosure. Since our launch, hundreds of other institutions have adopted this same tool without clear guidelines on transparency regarding the disclosure of AI use to patients. We share the subsequent experience and call for professional organizations to help create disclosure guidelines that can be used nationally.