Research conducted at UC San Diego Health on the types of editing that physicians across various specialties made to AI-generated drafts of patient messages provides some early evidence about physicians’ roles as humans-in-the-loop and offers some direction for AI optimization efforts.
