When the Eye Doctor Gets It Wrong: Using AI to Check a Glasses Prescription

An ophthalmologist ran a full exam and prescribed glasses for distance vision, and they work perfectly. But when the author put them on for computer work, focusing on the screen turned out to be impossible. Something in the prescription was off. An AI helped make sense of all the numbers.

What was wrong with the prescription

A typical glasses prescription looks simple: diopters (lens power), astigmatism (if any), and axis. But the ophthalmologist had literally mixed up two correction modes. The distance prescription was right, but the doctor got the near-vision parameters wrong (computer distance is about 60 centimeters).

The author noticed this after borrowing glasses from a friend of the same age with similar vision. The friend's glasses worked perfectly. That meant one thing: the doctor had made no mistake in the friend's prescription, but had made one in the author's.

AI as an analysis assistant

This is where a language assistant came in. The author simply wrote: "Here is my prescription, here is how the glasses perform, here is the doctor's diagnosis. Help me work out what's wrong." The AI went through every number and checked whether the combination of parameters made sense for a person of that age and with that vision history.

The result: the AI identified a clear inconsistency. The near-vision parameters were set as if for a person with a completely different refraction. When the author showed this analysis to the doctor, the ophthalmologist agreed that the error was his.

• The doctor did not check the prescription for both distances at the same time
• The AI analyzed the data without emotion or bias
• The doctor wrote the new prescription more carefully

When doctors make mistakes

Medical errors are not rare, but ophthalmology seems like a field where everything is simple: look into the device, prescribe glasses. In reality, an ophthalmologist sees hundreds of patients a week and fills out paperwork quickly. Prescription errors are infrequent, but they do happen. Usually they are caught on the spot: the patient says "I can't see with this," and the doctor redoes it. But sometimes the error only shows up at home, a week later, and then a second sign-off is needed.

What this means

The story shows that AI can be useful as a checking tool, not in a medical sense but in a logistical one. When someone needs to make sense of numbers, parameters, or data, the technology can help structure the information and point out inconsistencies. That lets a patient talk to the doctor in a more informed way. AI does not replace the doctor, but it can be a second pair of eyes when analyzing medical documents.