EP 23 Dr. Benjamin Slovis, MD, MA, Temple Health
Benjamin Slovis, MD, MA, joins STAT AI to discuss what it takes to move hospital AI from a promising pilot into everyday practice. Temple Health's Chief Medical Information Officer explores EHR integration, ambient documentation, clinician oversight, and the questions hospitals should ask vendors. The conversation also covers ethics, local validation, and the changing role of AI in medical training. The hosts open with Project Glasswing and the implications of AI cybersecurity for health care.
EP 20 Dr. Michael Prats, MD, and Dr. Shawn Sethi, DO
Michael Prats, MD, and Shawn Sethi, DO, join STAT AI to explore how AI is changing point-of-care ultrasound. The conversation separates automated measurements from image interpretation, examines what learners need to know before trusting an output, and considers the role of clinician judgment when a model disagrees. The guests also discuss access, infrastructure, external validation, and the workflow changes they expect next. The hosts open with a discussion of AI agents and OpenClaw.
EP 19 Dr. Resa E. Lewiss, MD, Brown University
Resa E. Lewiss, MD, joins STAT AI to explore how small, deliberate actions build expertise in ultrasound, AI, and clinical practice. Drawing on her experience as an educator and coauthor of MicroSkills, she discusses learning without embarrassment, including patients and caregivers in design, and assessing whether a tool actually saves time and energy. DJ and Ethan also consider patient-facing AI and the launch of ChatGPT Health. The episode offers a practical approach to experimenting thoughtfully and recognizing when to move on.
EP15 Dr. Andrew Taylor, MD MHS
In Episode 15 of the STAT AI Podcast, Dr. DJ Apakama and Dr. Ethan Abbott welcome Dr. Andrew Taylor, Professor of Emergency Medicine and Vice Chair for Research and Innovation at the University of Virginia. A national leader in clinical informatics, Dr. Taylor shares insights from two groundbreaking areas of his work: ECG foundation models and human-in-the-loop triage AI.
From training a model on over 10 million ECGs to implementing AI-assisted triage with nurse override across multiple hospitals, Dr. Taylor highlights both the promise and the pitfalls of deploying AI in real-world emergency departments. He and the hosts dig into automation bias, clinical trust, and what it means to balance algorithmic recommendations with human judgment at the bedside.
Together, they explore how foundation models could unlock new diagnostic possibilities, why outcomes—not just accuracy scores—should drive adoption, and how to prepare the next generation of residents to use AI as a partner rather than a crutch.
Tune in to hear how thoughtful, human-centered design is shaping the future of equitable AI in emergency medicine.