Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
I just celebrated a big milestone I hope I never reach: I maxed out my health insurance plan’s $10,150 out-of-pocket in less than five months of 2026, thanks mostly to two major eye surgeries. This means no more copays or coinsurance for in-network authorized care this year, as long as I continue to pay my monthly premiums.
But earlier this year, as I racked up what seemed like an eternal fountain of medical expenses, I couldn’t help but wonder if I was paying bills with errors. As a certified financial planner and longtime personal finance writer and editor, I’m aware of how many medical bills contain errors that make them more expensive.
Sometimes, medical bills contain obvious errors, such as charges for treatment that you explicitly declined. Otherwise, these errors are often difficult for the average patient to detect. Finding errors in billing can require clinical knowledge, along with an understanding of medical coding, revenue cycle management, and the arcane American health insurance system.
You may also have to sift through vast amounts of information. For example, I discovered that I had 87 insurance claims during the first four and a half months of 2026 and that the contract I signed during open enrollment was 149 pages long.
I had no desire to get an education in medical coding or ponder the meaning of 149 pages of insurance terminology, but I thought it might be generative artificial intelligence He will be up to the task. After all, AI excels at absorbing complex information and finding irregularities in huge amounts of data.
However, it turns out that using AI to find errors buried in my piles of medical bills wasn’t as easy as I had hoped. Here’s how I did it — and what I learned.
How I used artificial intelligence to find medical billing errors
I expected to find a plethora of AI tools to help patients identify billing errors. mistake.
Most AI tools aimed at improving billing accuracy are designed for providers, not patients, for obvious reasons.
The few patient-facing tools available often target a fairly narrow segment of billing issues. For example, Anti health It uses artificial intelligence to analyze bills and medical records to help patients understand why their insurance claims were denied and formulate an appeal. But there are a few patient AI resources that offer a general audit of your medical bills.
So I settled on usage Generative artificial intelligence – Specifically, $20 per month ChatGPT Plus subscriptionwhich was very helpful for me in drafting scripts to use with my insurance company when I tried to deny care.
My step by step process:
- I narrowed my focus to claims where I spent at least $150 to simplify the review.
- I retrieved the 146-page insurance contract and explanation of benefits, or EOBs, from my insurance company’s website.
- required Detailed medical bills of service providers, which is necessary to determine costs and inaccuracies.
- I compiled 14 itemized invoices and EOBs, as well as a spreadsheet summarizing all 87 of my claims.
- All personal information — such as my name, date of birth, address, and insurance ID number — was redacted from the documents before being uploaded to the AI.
Then I used the following ChatGPT prompt:
Serve as a medical billing expert and auditor with in-depth knowledge of the US healthcare system, medical billing codes, surgical billing practices, and outpatient billing practices. I will provide my insurance contract, an itemized bill and an explanation of benefits. Look for incorrect charges, unusually expensive or questionable charges, mathematical errors, charges that appear inconsistent with my insurance contract, and other possible errors.
Does ChatGPT detect errors in medical billing?
Before I even uploaded my itemized bills to ChatGPT, I could see an obvious flaw: How was the AI supposed to know if the bill accurately reflected the care I received?
For example, the first two itemized bills from the surgical center included 31 to 60 minutes of operating room time. But I didn’t bring a stopwatch with me to the surgery.
ChatGPT might have flagged it if I was billed for several hours of surgical time for a procedure that normally takes a few minutes. But how can ChatGPT know if, say, you’ve only been in the operating room for 28 minutes? Or whether the 200 drops I received before surgery were accurately reflected in the itemized surgery bill?
Instead, ChatGPT continued to focus on things like the fact that the amount my insurance company paid seemed ridiculously low compared to what the surgeon, anesthesiologist, and facility actually paid. Fair enough, but this is more an indictment of the lack of transparency of the American health care system than a sign of a billing error.
Amnesty International asked me to look at the one claim marked “rejected” in the spreadsheet. But the reason for the denial was that the surgeon had voluntarily withdrawn it and resubmitted it before the insurance company processed it. Some pharmacy claims were canceled, but they also had an easy explanation: The pharmacy automatically processed a number of refills I didn’t need.
I quickly lost hope that AI would help me find potential billing errors that I hadn’t previously identified. So I started asking honest questions about specific allegations.
There was one potential error I had already noticed: For one procedure, I was charged a $100 copayment for the specialist and a $150 copayment for a physician-administered medication, or $250 total. I spoke with an online customer service representative who said I should have only been charged for one service. So, I uploaded my live chat with the company representative, and asked him:
This conversation with an insurance representative indicated that I would only owe a retina specialist co-pay of $100 or a co-pay for doctor-administered medications of no more than $150 for anti-VEGF injections, but I was charged $250 for the visit and injection. Is this wrong?
ChatGPT quickly dashed my hopes on this front, directing me to the 149-page section of my insurance contract that states I am responsible for both co-payments. Clearly the insurance representative was wrong.
OK, but why did I pay $11,512 in copayments and coinsurance when my maximum liability to a patient was $10,150?
ChatGPT kept insisting that I only paid $10,150. Then it hit me: ChatGPT showed that I only paid $10,150 because that was my responsibility as a patient, according to my EOB.
Three weeks later, I had the same surgery on my right eye. Since I met my deductible, I had to pay less: $1,552, which I assumed was 50% coinsurance. But my EOB listed my patient liability at $999.
Again, I asked ChatGPT about this discrepancy. This time, she pointed out something that seems obvious in retrospect.
The $1,552 I paid up front was the amount I was actually responsible for after the first surgery. Since I was having the same surgery on the other eye, the facility estimated how much I would owe based on the first surgery, without taking into account how the patient’s liability would change after I met my deductible.
So ChatGPT confirmed that I overpaid $1512 for my second eye surgery, and it helped me understand why. But she couldn’t actually find the $1,512 overpayment on her own. I found this by keeping careful records of all the medical expenses I incurred.
What the AI flagged as potential errors
undefined
| index | Next step |
|---|---|
| Duplicate charges | Compare the terms with your EOB to ensure that the service was billed twice. |
| Denial or “Not Included” status. | Contact your insurance provider to understand the reason (coding error, missing information, or lack of authorization). |
| Fees for services not received | Review clinical notes or records and contact the billing department for a detailed explanation. |
| Mathematical errors | Add individual costs to ensure the final bill total is accurate. |
| Out-of-network fees for in-network care | Check your insurance contract and provider status list; Contact the facility to correct the billing category. |
Just providing ChatGPT with all the information needed to confirm the error requires a significant amount of work. In this regard, it seems as if using ChatGPT to audit medical bills is somewhat similar Tax filing software: It’s only as accurate as the data you provide, and collecting all of that takes a lot of work.
It is possible that my itemized medical bills contain additional errors. If they do, it’s an issue the providers and my insurance company will have to fight over. As long as I don’t have to pay more than the $10,150 out-of-pocket maximum — and I have no doubt that the amount I’m responsible for as a patient has reached that amount — I honestly don’t care if they have to fight among themselves; This is not my problem.
As of this writing, I’m still waiting for my $1,512 refund.
How to do this yourself
If you want to use AI to help you review your medical bills, keep these basic requirements in mind:
- Request detailed invoices: You are entitled to receive details of all costs incurred during the procedure. Contact your service provider to request this, as it is necessary to identify certain billing inaccuracies.
- Redaction of sensitive data: Before uploading any documents to the AI tool, remove all personal information, such as your name, date of birth, address, and insurance ID number.
- Maintain a personal spreadsheet: AI is only as accurate as the data it provides. Keep a detailed record of each medical claim, the amount billed, the amount your insurance company paid and the actual out-of-pocket payments. This manual tracking is crucial for spotting discrepancies between what you were charged and what you were actually responsible for.
