Skip to Main Content
Q+A

How Ambient AI Scribes Impact Medical Students

Jaideep Talwalkar, MD, discusses the impact of AI scribes on medical students and offers advice on how best to use this new technology.

5 Minute Read

As technology in medicine rapidly evolves, Jaideep Talwalkar, MD, is helping Yale School of Medicine lead efforts to integrate these new tools into undergraduate, graduate, and continuing medical education. His recent research focuses on the impact of ambient AI scribes on medical students.

“Having familiarity with a transformative technology like this is important for all of us so that we can have an opinion and a voice at the table,” says Talwalkar, professor of medicine (general medicine) and of pediatrics and associate dean for educational technology and innovation. “To speak about the cautions of AI, we need to have a firsthand understanding through using it.”

In a Q&A, Talwalkar discusses the benefits and risks of AI scribes for medical students and offers advice on how best to use this new technology.

What are ambient AI scribes and how are they used in medicine?

Ambient AI scribes are artificial intelligence tools that listen in the background to capture and transcribe conversations. This technology is used in clinical settings to help clinicians write their notes. A clinician has a conversation with a patient, and the ambient AI scribe can create a transcript and structured note in the patient’s electronic health record.

In the past three years, this technology has transformed the way many clinicians write their notes when seeing patients. Ambient AI scribes can help reduce physician burnout by reducing the after-hours time they spend on documentation and help patients feel like their doctors are more present and less on the computer.

What has your research suggested about the benefits and risks of ambient AI scribes when used by medical students?

Releasing this new technology to medical students early in their training is not without risk of harming their educational progress. At Yale, we have guardrails in place in that students write their own note first without the use of AI. Only after students write their notes do we release the AI scribe's output to them. Then the students submit a third note, a hybrid note, in which they reflect on their own note and the ambient AI scribe's note. Our recent study found no degradation in quality in the hybrid notes compared with the notes the students wrote on their own. That gave us confidence that we weren't causing harm by releasing this technology to students in a very controlled manner.

In another recent study, we had students work through two separate standardized patient cases, or clinical scenarios. In one case, students wrote their own note without the ambient scribe's input. In the other, they started with the ambient scribe's note and then submitted a hybrid note. When we compared the quality of the human note that they wrote in the first case to the hybrid note in the second, we found that there was some degradation in the quality of their writing in the parts of the note where they highlight their clinical reasoning—the assessment section and the plan section—after they had seen the ambient scribe's note. That's worrisome because it suggests that when students receive the tool's output before writing their own note, they may not think as much.

But it is important to expose medical students to ambient AI scribes to highlight potential problems early on because this technology is here, and they are going to be using it in practice. It's a double-edged sword.

What are other concerns about AI scribes?

A lot has been written about deskilling and never-skilling. My counter to that is that writing the note is just a means to an end. What we really care about is the student's ability to think. The goal is to come up with ways to expose students to this technology while preserving their ability to think through a problem after they've seen a patient.

My team is currently working on a custom-built ambient AI scribe for educational clinical workflows in which we don't release the assessment and plan sections to students. So the tool doesn't give them the parts of the note where clinical reasoning and thinking typically happen.

What advice about AI would you give to medical students, early career clinicians, and others who want to develop their clinical skills and grow professionally?

One fundamental principle is that we should use these tools to bring us closer to the patient, not create more separation. One of the success stories of the ambient AI scribes in clinical use is exactly that. As a physician, I am no longer sitting there with my laptop on my lap. I can have more eye contact, listen better, and be more present.

The second piece of advice is to encourage people to dive into this technology and to feel a sense of responsibility to do so, because we've all signed up for lifelong learning, and this technology is part of the landscape. In educational settings, we will soon be working with junior colleagues and trainees who are AI natives.

The more that the typical clinician or educator has some basic understanding of the technology, the more eyes we have on the possibility of unintended consequences.

General Internal Medicine, one of 10 sections in the Yale Department of Internal Medicine, is committed to the core missions of patient care, research, education, and community health from the “generalist” perspective. To learn more, visit General Internal Medicine.

Article outro

Author

Serena Crawford
Associate Director, Communications

Talwalkar J, Wright D, Schwamm L, Leydon G, Shabanova V
Impact of an Ambient AI Scribe on Medical Student Objective Structured Clinical Examination Notes: Nonrandomized Clinical Trial
JMIR Med Educ 2026;12:e88264
URL: https://mededu.jmir.org/2026/1/e88264
DOI: 10.2196/88264

Tags

Media Contact

For media inquiries, please contact us.

Learn more about the Yale Department of Internal Medicine

Visit the department website

Explore More

Featured in this article

Related Links