Joseph Kvedar, MD

A trial of 252 working adults found that a brief exercise about mental health being improvable helped those with poorer well-being return to a self-help app. A promising way to help people take the first step toward support. nature.com/articles/s41746-0…

"What the tool knows is non-contextual." Luciana D'Adderio on why patient-facing AI can help people feel included in their care, but expertise still sits with the doctor who can apply it to one specific patient. Full episode 👇

AI does not automatically remove bias from health care. In an npj Digital Medicine study using past ER records, patient-history details sometimes shifted AI priority away from patients who later worsened. Safer AI must consider these blind spots. newsweek.com/ai-could-push-h…

We just had a great conversation with Dr. Nigam Shah about Stanford's experience implementing large language models directly into their electronic medical record. You can watch the conversation below! cassyni.com/events/FXDprRt7C…

This week on The Digital Pulse, we sat down with @DBatesSafety and Luciana D’Adderio to ask how AI is transforming diagnosis. We got into why AI may reverse the traditional diagnostic process, whether clinicians can resist automation bias when the AI answer arrives first, what trainees lose if they never learn on legacy imaging, and how patients using AI at home are changing the doctor-patient relationship. And whether the doctor’s job is to make the diagnosis—or verify it. Timestamps: 00:00 Intro 01:43 How AI is changing diagnosis 02:49 Studying stroke AI in Scotland 05:20 AI reverses the diagnostic process 07:33 How stroke AI reads the scan 11:21 Speeding care in rural Scotland 11:54 A faster pace of change than ever before 13:49 AI in its real context of use 17:39 Training in the AI era 18:09 Confirmation bias and diagnostic error 22:13 Verifying the AI output 24:11 Automation bias and “don’t believe the beautiful picture” 26:40 What junior clinicians could lose 27:35 Patients using AI at home 29:20 The physician as advisor 31:00 Health anxiety and somatization 33:00 “Vibe diagnosing” and the last mile 34:48 AI is too eager to please 35:37 AI in mammography 37:28 A broader revolution in diagnosis Hosted by @BenMazer and @sammargolis_. Produced by @danielkimjung.

This week on The Digital Pulse, we sat down with @tdazad, @suchisaria, and @HealthPrivacy to ask what Henrietta Lacks can teach us about medical AI. We got into why HIPAA may not be enough for the era of generative AI, whether meaningful informed consent is possible at the scale modern AI requires, what patients should know about where their data goes, and who should decide when health data can be commercialized. And whether the answer is more consent—or more transparency. Timestamps: 00:00 Intro 01:22 Ethics and privacy concerns 02:49 What Henrietta Lacks can teach us about AI 06:21 Is HIPAA enough? 10:14 Data brokers and where health data comes from 11:30 Patient data and AI 15:35 The limits of informed consent 17:15 What real data transparency could look like 22:52 Can transparency work at the bedside? 25:05 Commercialization and ethics boards 28:51 Independent vs. university ethics boards 32:28 Building a clinical intelligence platform 37:03 Tech mindset vs. regulation 40:00 The cost of bureaucracy 42:11 Closing thoughts

Tune in for the live recording of our latest podcast tomorrow, Friday August 21 at 3:30 PM. We will discuss what the Henrietta Lacks story can teach us about the ethics of medical AI cassyni.com/events/2vDLEJ8AA…

A Stanford survey asked 1,871 US adults whether they use chatbots for their mental health. 24% said yes. Rescaled against national LLM use rates, the authors estimate 14 to 18 million American adults.