The Power of AI in Otolaryngology

This is an exciting time in healthcare. Artificial intel­ligence (AI) is beginning to show its potential not only to improve patient care but also to trans­form the provider and staff experience. For physicians, one of the most important questions to ask is: What parts of our jobs do we like least, and how might AI help us accomplish those tasks more efficiently?

At its best, AI can free up precious time—allowing surgeons to spend more time with patients and more time thinking, rather than typing. I have previously written about the importance of making healthcare jobs into “good jobs.” AI offers the potential to make jobs better by making administrative work less laborious and more accurate. AI has the potential to increase access to care and drive patient-centric, individualized care.

One of the clearest examples is the AI-powered scribe. Many in our field have been experimenting with these tools for several years. My own institution recently expanded access, and for the past few months, I’ve been using one to assist with my clinical documentation. I am in love with the time it saves and the added detail in my notes. While I still don’t finish every note by the end of clinic, the hours I used to spend typing late at night have been dramatically reduced.

This is not my first experience with scribes. A decade ago, I tried using in-person scribes, but the model was imperfect. Patients disliked having another person in the exam room. I felt responsible for making sure the scribe had meaningful work and adequate space. Turnover was frequent, and retraining was a constant burden. Unfortunately, human scribes made more work for me and not less. Ultimately, my patients and my team told me what I already knew: Scribes weren’t working in my practice.

AI scribes are different. They do require some initial training, but they don’t leave for medical school, and my patients haven’t objected to their presence. I find I can give my full attention to patient conversations because I’m not distracted by note-taking. For me, the result is better focus during visits and more complete documentation afterward.

Fifteen years ago, tools like REDCap felt like the future, with the promise of integrating structured data into the electronic health record. But the integration never materialized, and duplicative data entry proved unsustainable. Large language models offer a fundamentally different path. They can extract key elements of the exam without requiring rigid, “databaseable” fields. Of course, this makes it more important than ever that information entered into the record is accurate—because the AI can only be as precise as the data it receives.

Beyond documentation, I find inspiration in working with startups developing these technologies. Their passion and urgency are invigorating, a refreshing contrast to the often-slow pace of academic medicine. It reminds me of the entrepreneurial energy I valued during my executive MBA training—the sense that obstacles are not immovable walls but puzzles that can be solved with creativity.

AI in otolaryngology is still in its early stages, but its potential is enormous. Used thoughtfully, it can reduce burdensome tasks, restore time for patients, and reenergize clinicians and staff. For me, it has already done all three. (For this editorial only, I used AI to reduce the length to 500 words and provide editorial comments after my draft—how did it do?)