Oh, what a time to be alive! In an era where we can ask machines to diagnose our itchy eyes, a groundbreaking study decided to put ChatGPT to the test with the riveting topic of uveitis. Because, you know, why ask a real doctor when you can chat with a bot?

So, they rounded up 32 burning questions about uveitis, crafted by none other than a uveitis specialist (because who else knows how to stump a chatbot?). These questions were then fed to ChatGPT 3.5, which, in its infinite wisdom, churned out answers that were compiled into a survey. This survey was then handed over to five uveitis specialists, because apparently, they had nothing better to do than rate chatbot responses on a scale of “more correct than incorrect” to “nearly all correct”.

The verdict? ChatGPT scored a median accuracy of 4.00, which is somewhere between “Hey, it’s not completely wrong” and “Wow, it almost sounds like it knows what it’s talking about”. But when it came to completeness, oh boy, it scored a whopping 2.00, which is the academic equivalent of “Well, at least it tried”.

And let’s not forget the interrater variability assessment, with kappa values of 0.0278 for accuracy and 0.0847 for completeness. If those numbers don’t scream “consistency”, I don’t know what does.

In conclusion, while ChatGPT might give you a somewhat accurate idea of what uveitis is, don’t throw away your medical textbooks just yet. It seems our AI friend still has a long way to go before it can replace your eye doctor. But hey, at least it’s trying, right?