Clinical AI has moved from pilot projects into everyday care. According to the AMA, 81% of physicians now use AI in their professional work, more than double the share in 2023 (AMA, 2026). The question is no longer whether AI belongs in medicine. It’s how we use it well.
At Aidoc, we are shaping the future of care with clinical AI. That means making room for honest conversations about what AI changes for patients, clinicians and the health systems that support them. That’s why we’re launching AI on the Front Lines, a new podcast about the real-world decisions behind clinical AI.
The show is hosted by Dr. Jesse Ehrenfeld, Aidoc’s Global Chief Medical Officer. Jesse is a practicing anesthesiologist, a former president of the American Medical Association and a former Commander in the U.S. Navy. Few people have seen healthcare from as many angles: at the bedside, in national policy debates and inside the technology itself. Every other week, he brings a new guest to the mic, from clinicians and health system executives to policy experts, for a candid look at what’s working, what isn’t and what comes next when it comes to AI in healthcare.
In Episode 1, Jesse sits down with Dr. Céline Gounder to explore:
- Why clinicians must still be able to explain their decisions as AI in healthcare becomes routine
- How a careful patient history helps clinicians spot when an AI recommendation doesn’t fit
- How medical skilling should change in the age of AI
- Tiered care, prior authorization, and whether Céline would choose medicine again
Episode 1: Dr. Céline Gounder
Jesse’s first guest is Dr. Céline Gounder, an infectious disease specialist at Bellevue Hospital and NYU and medical correspondent for CBS News. The two trained together at Massachusetts General Hospital, and their shared history makes for a frank, wide-ranging conversation.
They cover a lot of ground, from which specialties AI is likely to reshape to how it could change access to care. At the center of it all is a challenge every health system will face: preserving clinical judgment as physicians come to rely on AI in their daily work.
Preserving clinical judgement when using AI
Jesse describes a situation familiar to anyone who trains clinicians. A patient’s blood pressure falls in the OR, an algorithm recommends a treatment, and the resident follows it. The choice is probably correct. “But when you ask them, like, why was that the right thing to do… they can’t explain it,” he says. “And that’s where I get a little uneasy.”
For Céline, the remedy starts with a skill as old as medicine: a careful patient history. Where has the patient traveled? What makes this case unusual? Without that context, a clinician can’t tell whether an AI suggestion is a smart catch or a poor fit. The history is what lets a physician check the machine rather than simply defer to it.
This isn’t just a theoretical worry. In the same AMA survey, 88% of physicians said they have at least some concern about losing skills because of AI. Seventy percent were worried about skill loss among today’s medical students and residents (AMA, 2026).
Rethinking how clinicians build skills
If AI is changing what clinicians do every day, Céline argues, it should also change how they learn. She shares several ideas in the episode:
- Keep practicing without the tools. Systems go down, whether from a power failure, a software upgrade or a cyberattack. “I think they need to be able to do some of this without the tools,” she says. “Otherwise you’re going to have significant deskilling of the workforce.”
- Learn AI, including its limits. Future physicians should understand how to use AI and where it falls short, not just how to avoid it.
- Make room for what AI can’t do, like taking a good history, communicating with empathy and earning patients’ trust.
- Add new skills to the curriculum. She points to AI ethics, health policy and healthcare financing as gaps in training today.
- Stay adaptable. With roles changing quickly, she encourages young clinicians to build skills outside their specialty so they can move between careers over a lifetime.
Jesse adds a habit from his own practice: deliberately rotating between tools, including ones he’s less comfortable with, so his skills stay sharp for the patient who needs them. The two also discuss whether that kind of deliberate practice can survive pressure to see more patients.
Tiered care, and beyond
The episode goes well beyond training. Other highlights include:
- AI that saves time when every minute counts. Jesse shares a recent case from his own practice: a stroke patient whose condition was flagged by an Aidoc algorithm running in the emergency department. She was rushed to have the clot removed and went on to do well.
- Different patients, different kinds of care? The AAMC projects a shortage of up to 86,000 physicians in the U.S. by 2036 (AAMC, 2024). AI could help each clinician reach far more patients. Céline explains why that could lead to different levels of care for different people, and what it might mean for rural communities.
- AI on both sides of prior authorization. 61% of physicians worry that health plans’ use of AI is driving more prior authorization denials (AMA, 2025). Céline describes what that looks like on the hospital wards, and how new tools are emerging to help patients push back.
Jesse also puts a personal question to Céline: knowing everything she knows now, would she go into medicine again? It’s a candid moment you’ll want to hear for yourself.
Listen now
The first episode of AI on the Front Lines is available now. Listen to the full episode on YouTube, Spotify or Apple.
New episodes come out every other week. Follow along on social media and subscribe to the newsletter to hear about upcoming guests. Each guest leaves a question for the next one, and Céline set a high bar: How do we incentivize ethical, equitable use of AI in medicine?
Sources
