Chatbots provided incorrect, conflicting medical advice, researchers found: “Despite all the hype, AI just isn’t ready to take on the role of the physician.”
“In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice,” the study’s authors wrote. “One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care.”
Just sharing my personal experience with this:
I used Gemini multiple times and it worked great. I have some weird symptoms that I described to Gemini, and it came up with a few possibilities, most likely being “Superior Canal Dehiscence Syndrome”.
My doctor had never heard of it, and only through showing them the articles Gemini linked as sources, would my doctor even consider allowing a CT scan.
Turns out Gemini was right.
It’s totally not impossible, just not a good idea in a vaccuum.
AI is your Aunt Marge. She’s heard a LOT of scuttlebut. Now, not all scuttlebut is fake news, in fact most of it is rooted at least loosely in truth. But she’s not taking the information from just the doctors, she’s talking to everyone. If you ask Aunt Marge about your symptoms, and she happes to have heard a bit about it from her friend that was diagnosed, you’re gold and the info you got is great. This is not at all impossible. 40:60 or 60:40 territory. But, you also can’t just trust Marge, because she listens to a LOT of people, and some of those are conspiracy theorists.
What you did is proper. You asked the void, the void answered. You looked it up, it seemed solid, you asked a professional.
This is AI as it should be. Trust with verification only.
congrats on getting diagnosed.