TLDR: A new article in JAMA, co-authored by Stanford’s Michelle M. Mello, outlines a framework for when patients should be informed or provide consent regarding the use of AI tools in their medical care. The guidance addresses the diverse applications of AI in healthcare, from surgical robots to administrative tasks, emphasizing a nuanced approach to patient notification.
As artificial intelligence (AI) increasingly integrates into healthcare, a critical question arises: when and how should patients be informed about its use? Michelle M. Mello, a distinguished professor of law and health policy at Stanford University, along with co-authors Danton Char, a pediatric cardiac anesthesiologist, and Stanford student Sonnet H. Xu, have addressed this complex issue in a recent article published in JAMA. Their work, titled ‘Ethical Obligations to Inform Patients About Use of AI Tools,’ provides a comprehensive framework to guide healthcare providers on patient notification and consent for AI applications.
The article, published on July 29, 2025, argues against a one-size-fits-all approach to AI disclosure due to the vast array of AI tools currently employed in healthcare. For instance, an AI-guided surgical robot, which carries considerable risk and offers patients the option of non-robotic surgery, would necessitate direct patient consent. Conversely, a generative AI tool used by hospitals to streamline prior authorization requests with insurance companies would likely not require patient notification or direct consent, as it operates in the background with minimal direct patient impact.
Mello and her colleagues highlight that while some AI technologies, such as ambient listening tools, are overtly apparent to patients, many others function discreetly and may not be obvious during clinic visits. The framework aims to help healthcare leaders and clinicians navigate this ‘thorny terrain’ by providing clear guidelines for various use cases.
Recent surveys underscore the public’s apprehension regarding AI in healthcare. A significant 60% of U.S. adults reported discomfort with their physician relying on AI, and 70% to 80% held low expectations for AI to improve crucial aspects of their care. Furthermore, only one-third of respondents trusted healthcare systems to use AI responsibly, and a substantial 63% expressed a strong desire to be notified about AI use in their care. This data emphasizes the importance of transparent communication and a well-defined consent process.
Also Read:
- AI Agents Revolutionize Healthcare IT, Delivering Significant Time Savings
- Africa AI Health Forum Explores Generative AI’s Role in Revolutionizing Healthcare
The authors’ framework focuses on AI tools that involve human oversight, acknowledging that fully autonomous AI in healthcare remains rare. This nuanced perspective aims to balance the benefits of AI innovation with the ethical imperative of patient autonomy and informed decision-making.
