Bridget Duffy, MD has spent her career working at the crossroads of technology and humanity in healthcare. She helped lead the patient experience movement before most organizations paid attention to it, and she learned that you cannot improve the patient experience without first fixing the experience of the people delivering care. In this conversation from HMPS 2026 in Salt Lake City, she argues that AI can humanize healthcare when it eases the burden on burnt-out doctors and nurses instead of adding to it.
Her test for any AI tool is whether it removes friction from the patient journey and takes work off the clinical team. She points to AI triage in the call center, recovering patients who drop out of online scheduling, and emergency-room tools that update patients on wait times, translate lab results into plain language, and monitor progress so nurses can spend more of a 12-hour shift at the bedside.
Duffy is part of the team behind the AI Care Standard, a 10-pillar checklist developed through Vital.io that defines what has to be in place before a health system deploys AI: safety, trustworthiness, and sensitivity in how information reaches people. Her clearest example is how patients learn serious news. No one should find out they have cancer from an app or an automated message. The standard routes those results to a human, alerts the physician, and gives the clinician the chance to deliver difficult news in person.
She also presses on transparency. Patients increasingly trust AI without knowing what they are trusting, so any health system using these tools should disclose upfront when a patient is interacting with AI. Nursing has been the most trusted profession year after year, and Duffy believes AI can earn that kind of trust only with clear disclosure and a human in the loop.
The conversation closes on a lesson from one of her early mentors, a pulmonologist who asked his students how you know when to admit a child with pneumonia. They reached for the lab values and the chest X-ray. He told them you admit a child when they have lost the twinkle in their eye, a judgment no dataset captures. Medicine is an art and a science, and Duffy’s view is that AI can deliver the data while a human holds final say over decisions about someone’s loved one.

