I Asked ChatGPT to Help Me Figure Out My Mystery Illness

When I ask my 19-year-old sister if she uses AI in college, I’m only half-joking. To me, AI is less of a tangible tool and more of an abstract robot army to be avoided at all costs. Her bewildered “what are you, 90?!” response is a swift (and mildly embarrassing) reality check. Noting my confusion, she explains how free software like ChatGPT can answer almost any question and hold human-like conversations. Having dealt with an undiagnosed mystery illness for the past three years, I started thinking about the answers I want more than anything else in the world.

Within the realm of human intelligence, I’ve already been to five different hospitals, desperately trying to figure out the reason behind two ischemic strokes, and the loss of vision in my right eye. I’ve worked my way through treatments like corticosteroids, radiation therapy, spinal taps, angiograms, and biopsies, but nothing seems to stick. In terms of experts, I bounce from neurosurgeons to rheumatologists to oculoplastic specialists, hoping someone will be able to connect the dots. People say I just need to try a different hospital, a different doctor, a different specialist, but they don’t know what it feels like to constantly come up short.

Given that I’m still experiencing persistent flare-ups, the possibility of an enucleation (the removal of my eye), is slowly inching its way into reality. The fact that a software with the ability to instantly assess almost 570GB of data exists, why wouldn’t I give it a shot?

My Experience With ChatGPT

My initial impression of ChatGPT is that I’m bad at texting robots. I have a lot of questions, but I worry about overwhelming the AI, my fingers awkwardly hovering over my keyboard. I start by punching in “medical diagnosis” and, in this moment, I do feel 90. ChatGPT promptly informs me that — while it can offer clarity on certain symptoms and conditions — it is not a doctor, nor is it a replacement for professional medical advice. “If there’s something specific you’d like to know or discuss, feel free to share, and I’ll do my best to assist you!” it responds, and I decide to take its cheerful advice.

“What are some reasons for ocular inflammation?” I ask. The answers are detailed, listing common causes like infections, allergies, and physical injuries to the eye, but are relatively unhelpful for my specific circumstance. Realizing ChatGPT has no way of knowing anything about my medical history, I decide to give the robots more context to work with.

“Are there any autoimmune conditions that may cause nongranulomatous inflammation in the orbit, cutting off blood flow to the optic nerve?” I type into the chat bar. It’s such a specific query, I almost feel stupid writing it. Luckily, ChatGPT thrives on specificity. “Autoimmune conditions can indeed cause inflammation in the orbit (the area surrounding the eye), leading to various eye-related symptoms, including optic nerve ischemia,” the chatbot replies, referencing the process that occurs when the optic nerve does not receive enough blood flow. It lists conditions like rheumatoid arthritis, systemic lupus erythematosus, and thyroid eye disease (aka Graves’ orbitopathy) as examples. Although I’ve tested negative for most of them, I’m encouraged by the amount of detail.

My next question has more to do with treatment. “Is there a way to help nongranulomatous inflammation in the orbit . . . if Rituximab and corticosteroids don’t work?” I ask, referencing my previous medications. I’m curious to see how the chatbot will respond to this stipulation, but it takes the challenge in stride. While some specialists have shrugged and made me feel like there were no other options available, ChatGPT creates an extensive list of potential treatments in seconds — a few of which I’ve heard before, and a few I haven’t.

The bullet points encapsulate everything from specific TNF inhibitors to plasma exchange, adding that surgical invention (like the kind I’m currently considering) is also a very real possibility. Given that my condition is so difficult to research in any meaningful way, this information feels strangely validating. I wouldn’t let AI dictate my decision around removing an eye, but I appreciate its input nonetheless.

Speaking of which, it’s time to see how ChatGPT feels about enucleation. I ask if it’s possible to remove an eyeball if you have a mass behind it, including a few details about my circumstances. For the first time, the robots seem taken aback. This is a drastic surgery, ChatGPT tells me, but it may be necessary for a few reasons. The bot lists them out in a way I feel would be quite helpful to anyone trying to make sense of a very difficult situation (myself included). I ask a few follow-ups about the healing process, and notice that in each of its answers, ChatGPT emphasizes the emotional toll a surgery like this might have on a patient. I’m ashamed to say I hadn’t expected so much thoughtfulness from AI, but I appreciated the virtual concern for my emotional and psychological well-being, even if only via chat.

Did ChatGPT Solve My Mystery Illness?

Compared to a clinical setting where you only get a few minutes to ask all your most critical questions, ChatGPT gives you time and space, free from any in-person judgment. Although AI isn’t as thorough or personalized as information from a healthcare professional (and we still don’t recommend using it as a replacement for actual medical care), it may help you uncover the questions you should be asking. Are there treatments you haven’t considered? Is there a specialist that could offer additional insight?

I’ll admit it’s slightly emotional that even the robots don’t know what’s wrong with me. In a sad last-ditch effort, I seriously considered typing something like, “Why is this happening to me?” until I remembered an existential ask would get me nowhere.

While I still don’t have a diagnosis, I walked away from my first ChatGPT experience feeling hopeful, validated, and heard, which is more than I can say about some of my other appointments out in the real world. And even if the robots aren’t the definitive solution, at least they won’t send me another medical bill.

Chandler Plante is an assistant editor for POPSUGAR Health & Fitness. Previously, she worked as an editorial assistant for People magazine and contributed to Ladygunn, Millie, and Bustle Digital Group. In her free time, she overshares on the internet, creating content about chronic illness, beauty, and disability.