HAS THE BUZZ of artificial intelligence (AI) passed by or are you finding new ways to incorporate it into your practice? Much of what we dreamed possible with AI seems to still be in the future, but with every passing day practitioners are finding ways it can simplify and enhance clinical practice and business management.
AI has been shown to be helpful in analyzing topographies (Kading, 2024) and predicting ideal contact lens fittings (Kading, April 2025). Across other areas of health care, we continue to see advancements in practice diagnosis and treatment using AI. One area in which AI will be helpful for eyecare providers is the ability to assess our own clinical data findings when compared to large sets of data; we look forward to a time when AI can review millions of data points to predict outcomes based on different clinical findings. However, we may not see that benefit for years due to privacy and a lack of electronic health record interoperability (Kading, July/August 2025).
The question clinicians ask today is how to go about making AI a more impactful part of their everyday patient care. One approach is to use the large language models that are currently available with advanced chat features.
When encountering a challenging case that would normally drive a clinician to an internet search, a textbook, or the literature, consider using AI as your research assistant.
Using your smart phone or computer, capture with the camera or upload images that have had the patient’s protected health information (PHI) removed or cropped out. Next, explain what you want the model to do:
“Act as a top 1% expert in [contact lenses, cornea, glaucoma, etc], who has deep knowledge based on clinical experience and research. I’ll be presenting a case and want research-backed guidance; do not make assumptions. Ask clarifying questions and provide documented evidence when I ask for it. We will be working to confirm a diagnosis and identify proven treatments.”
Then, using the chat feature, explain what you are encountering.
“I am seeing a patient who has irritation of their eyes. They have circumlimbal injection. I am noting some white dots in the peripheral cornea that do not stain. I have uploaded images of what I am seeing. What are the differentials? What is the most likely diagnosis? What other tests would be helpful for me to perform? And what treatments should I consider?”
Another example:
“My patient is wearing a scleral lens that was fitted using profilometry. They are noting some discomfort upon blinking. I have uploaded their profilometry maps as well as optical coherence tomography images of the landing of the scleral lens. Help me differentiate between the surface being dry, which is causing discomfort, and an edge issue. As far as I can tell, the lens is wetting well, and the edges of the lens look ideal. Suggest other possible problems that I am missing.”
Although we are far more capable than AI, astute clinicians will use the tools at hand to optimize their patient outcomes. AI has a broad grasp of information on the internet, including case reports and research, and can also help to guide you through clinical encounters like a research assistant can. But you are still the clinician with your name on the chart. So, eyecare practitioners still need to test and retest anything AI suggests (just as they would do with an internet search).
Remember that AI can take any information found on the internet—including false information—when making its diagnosis, so clinical expertise is critical when considering the AI results in patient care. But when AI can supplement clinical knowledge with alternate perspectives, patients may benefit from better research-based outcomes.
References
1. Kading DL. AI & better lens fits. Contact Lens Spectrum. 2024;39(8):30. https://clspectrum.com/issues/2024/october/ai-in-practice
2. Kading DL. AI + first-fit success. Contact Lens Spectrum. 2025;40(3):28. https://www.clspectrum.com/issues/2025/april/ai-in-practice
3. Kading DL. AI’s frustration with the dry eye space. Contact Lens Spectrum. 2025;39(6):36. https://www.clspectrum.com/issues/2024/julyaugust/ai-in-practice
