Chris Spyrou
👤 SpeakerAppearances Over Time
Podcast Appearances
We know that ChatGPTC sees over 40 million health-related queries a day.
I don't know what portion of that we can attribute to your searches, however.
More and more people are sharing their details on there and they're sharing intimate and detailed symptoms because they prefer doing that with a chatbot over a doctor or a human.
Was that part of the appeal for you?
If not trying to, you know, avoid the embarrassment, but the idea of not wanting to pester medical professionals that you know you're going to be waiting a very, very long time to see and paying a lot of money to see?
Have you run into any issues with hallucinations and stuff like that?
Because we have seen Google recently make changes to its AI summary, which was lifting amateur medical advice, putting it up to the top and presenting it as fact, and then
a bunch of journos did some work and Google made some tweaks to that and things like that.
Have you come across anything like that where, you know, the chatbot is really pushing you in a direction you're like, oh gosh, no, no, no, I do not have this.
And similar to your story, when I was overseas last year with a group of friends, we had a mate that definitely broke his hand and we were like, you've broken your hand, chat GPT confirmed it.
But it had that little disclaimer down the bottom that was like, go see a doctor, dude, you probably need surgery.
And he did need surgery.
But interestingly, we also used it for other stuff to kind of dissect formal diagnoses, even trying to get the jump on an x-ray.
I mean, we've all been in the car after getting an x-ray and lifted it up to the sun and trying to make sense of where the fracture is and things like that.
I found that ChatGPT has been helpful in that space.
Is that sort of the way you're using it as well or not really?
Poppy, thank you so much for joining us to talk us through how you're using AI in healthcare.
It's fascinating stuff and it's a fascinating chat.
That was journalist Poppy Reid speaking with me there, and we reached out to the TGA about all of this, and their advice is to remember that general AI tools like ChatGPT are not considered medical devices, and so users run the risk of receiving unsafe and inaccurate information when it comes to their health.
They've also provided us a list of AI-enabled medical devices that are TGA-approved.