Everyone’s talking about AI in eye care—but how much of that talk is turning into actual practice?
To get a pulse check, Vidyat Udayasankar and I ran a short survey earlier this year with 22 eye care professionals—13 ophthalmologists and 9 optometrists—across the U.S. and abroad. It wasn’t meant to be exhaustive, but rather directional. And what we saw suggests that while interest in AI is high, real-world adoption is still in its early days.
Familiar, But Not Fluent
About two-thirds of respondents said they were “somewhat familiar” with AI in health care. That usually means they’ve heard about it, maybe seen a demo or read an article, but haven’t yet integrated it into day-to-day practice. Less than 10% felt “very familiar” or comfortable with AI tools. And 23% acknowledged little to no exposure.
That’s not surprising. Most of us weren’t trained to think about machine learning, image segmentation or natural language processing in school. We’re clinicians first. But as AI becomes more embedded in diagnostics, decision support and workflow optimization, we all have to start building some fluency—not just for ourselves, but for the patients asking questions we’ll need to answer clearly.
Early Use, Promising Signals
Roughly a quarter of respondents said they’re already using AI in some form—typically for image-based screening (retina, glaucoma), clinical decision support or streamlining surgical planning. Those who use AI tools tend to use them daily, and they report improvements in time management, diagnostic confidence and efficiency.
In short: early adopters are seeing real value. But they remain the minority.
What’s Holding People Back?
The barriers aren’t new:
- Poor integration with EHRs continues to frustrate clinicians.
- Lack of structured training leaves providers unsure how to vet and implement tools safely.
- Patient skepticism lingers, particularly around machine-generated diagnoses.
- Cost and infrastructure gaps make it harder for smaller practices to explore adoption.
All of these are solvable problems, but they require coordination between industry, providers and payors. We can’t just build better tech—we have to fit it into the clinical rhythm.
A Profession in Transition
Despite the challenges, 68% of respondents said they believe AI will be a standard part of eye care within five years. That sense of inevitability is encouraging. But as we saw in this small survey, belief doesn’t always equal readiness.
That’s one reason we created AI in Eye Care and our Real Talk podcast—short, accessible episodes that break down core concepts, real-world use cases, and what’s coming next. If you’re looking to stay oriented in this evolving space, we invite you to follow along. We’re all learning, and we’re better off doing it together.
The data from this survey doesn’t suggest that eye care is behind. Rather, it reflects a specialty actively negotiating how—and when—to bring AI into the fold. As the tools improve and the evidence mounts, so will confidence. And as clinicians, we’ll remain at the center of that process, making thoughtful decisions about how best to use the technology in service of our patients.
Let’s keep the conversation going.
Vidyut Udayasankar is a rising junior at Emerald High School with interests in medicine, AI and drug development. He is also a nationally ranked tennis player.
Authors
