By Thanh Mai, OD
March 19, 2025
An exciting new opportunity is here, though too few ODs are taking advantage of it: artificial intelligence.
We all have heard of it, of course, but how many of us are giving it a try in our own lives and in our offices?
You can start as small as ChatGPT, and move on from there to bigger and more impactful ways of optimizing this new and ever-evolving technology
How do I get started?!
I recommend starting with www.chatgpt.com. Go now if you never have and ask any question. Use ChatGPT instead of Google the next time you have a question.
Make sure you go to the upper right corner where your profile is on ChatGPT, and go to Settings. From settings go to Data Controls. I recommend turning off “improve the model for everyone.”
Think of AI as holding a giant sphere of everything known to humankind. When you are prompting it, the AI needs to know what part of its mega sphere to start. So, one of the best ways is to start with a prompt like: “You are an optometrist offering dry eye treatments for patients suffering from dry eyes.”
That tells the AI bot how to start thinking. You might wonder how much to charge for your new intense pulsed light (IPL) services. Here is a prompt you could use to help you answer that question:
“My offer is IPL and radio frequency (RF) to improve tear production and eliminate dry eye symptoms. My unique selling point is that we have a med-spa environment. Determine an optimal pricing strategy for my service using the value-based pricing approach, and give me a recommendation for which price I should charge for my services.”
You can swap out IPL/RF with vision therapy or whatever. But that would be a fantastic prompt to get to a reliable endpoint of what you want.
Remember also that you can upload files to ChatGPT and ask it for feedback. Examples of what I used recently include:
- Giving ChatGPT my lease agreement and reviewing any areas of concern.
- Helping a colleague write a business plan and a pro forma for a loan request
- Editing a CV or resume
Hallucinations: AI Needs to Be Double-Checked
ChatGPT and most AI is accurate 98-99 percent of the time. However, at times it aims to please no matter what. When it doesn’t know, it sometimes just makes things up. When it starts to spit out things that do not seem credible, we call this an AI hallucination. So, for now, I still consider AI like a summer intern.
Other Articles to Explore
It can help me get started writing a report or a doctor’s note for a patient. However, I certainly check it before signing off.
Your own Custom GPTs
Do you remember the days of the app store when the commercials were touting “Well, there’s an app for that!”? Well, in ChatGPT there is a GPT store. Perhaps in the future there will be “a GPT for that” commercial.
You can create your own AI model and teach it how you want it to behave. From teaching it how to bill medically necessary contact lenses to how you want your receptionist to answer the phone, you can customize your own AI agent.
Here is an example of one I made for this article. I uploaded essentially a ton of articles from Review of Optometric Business, threw in some sage advice from billionaires and created this GPT on how to scale your optometric business. Try it for free, no financial disclosures.
Harnessing the Power of AI to Improve Care
AI is primed to become the best technological friend of the average OD.
There are already FDA-approved use cases (IDx-DR for instance) for AI in the ophthalmic space. And it will only accelerate with time.
If you uploaded a posterior segment photo and your exam notes, an AI bot would likely be able to diagnose the posterior segment disease with surprising accuracy.
The consensus, though, is that AI will not replace anyone directly, but everyone will be replaced by someone who is a multiple more efficient because they use AI.
The Future of AI
There will come a time when the check marks you see on Instagram and other social media sites to note authorized accounts are not just for those who are famous.
There may be check marks to verify if an account you are interacting with is human or not. Because in the future, people will inadvertently be interacting with AI personas and have no idea.
Or on the flip side, they WILL know they are interacting with an AI persona, but may choose to do so for companionship and even therapy.
AI is not the future. It is already happening now. Get on board–and see how you, your practice and patients benefit in small and large ways.
Click HERE to check out the new ROB sister publication, AI in Eye Care.
Thanh Mai, OD, is an owner of Insight Vision Center Optometry, a Vision Source practice in Costa Mesa, Calif, Optometry Corner, a Vision Source practice in Irvine, Calif. and Eyecon Optometry, a Vision Source practice in in Reseda, Calif. To contact him: tmai@visionsource.com
