Health Advice is not all the same. EMERGENCY by Mary Fernando

  Worried about your health? Need some advice? What if you can access information right now? 

Interested?

What if I told you that the diagnosis is correct only about 37% of the time? Still interested? 

What if I told you the information may be a hallucination, so you can get recommendations to treat your illness by eating rocks? 

What if I told you that even with all these limitations, that (32%) of adults are turning here for health information and advice. - Interested? 

AHH you'd say; AI. 

Exactly. 

The problem is that patients get nervous, understandably it's their health, and may not get access to their doctor, if they're lucky enough to have one. So they turn to their phone or computer, ask a bunch of questions and get a bunch of answers. 

Many believe the AI is the new genius on the block. They've been told AI can process information faster than humans (true) even human doctors (also true) so what could go wrong with asking AI?

Let me count the ways. 

"One problem is that AI doesn't differentiate between accurate information and bad information, so when you're asking about your symptoms or a treatment for your diagnosis, "AI Overviews do quote sources like the World Health Organization, WebMD, PubMed, and Mayo Clinic. However, they are not limited to these sources. Analysts say that the tool pulls information from Reddit, Wikipedia, different blogs, and e-commerce sites." 

So, a blog by Bobby Wonder who is a mechanic (even a darn good one) who writes stuff down after a couple of beers and a bottle of rum, because he is inspired, is treated equally as the peer reviewed PubMed publication. Further, AI doesn't have a sense of humour, so even statire is treated as valid as medical publications, resulting in the documented advice to eat rocks. Daily. 

A University of Waterloo-led study found that ChatGPT-4o, the AI language model by OpenAI, answered nearly two-thirds of open-ended medical diagnostic questions incorrectly...Only 37% of ChatGPT’s answers were correct, with most deemed unclear by experts and non-experts alike. The study highlights the risks of relying on AI for self-diagnosis, as people may receive false reassurance or unnecessary alarm. 

Surely, someone can sort the wheat from the chaff in advice and, presumably that someone is you. 

MIT did a study on just that: 

"We conducted a study in which a total of 300 participants gave evaluations for medical responses that were either written by a medical doctor on an online health care platform or generated by a large language model and labeled by physicians as having high accuracy or low accuracy. Results showed that participants could not effectively distinguish between AI-generated responses and doctors’ responses and demonstrated a preference for AI-generated responses, rating high-accuracy AI-generated responses as significantly more valid, trustworthy, and complete/satisfactory. Low-accuracy AI-generated responses on average performed very similarly to doctors’ responses. Participants not only found these low-accuracy AI-generated responses to be valid, trustworthy, and complete/satisfactory, but also indicated a high tendency to follow the potentially harmful medical advice and incorrectly seek unnecessary medical attention as a result of the response provided. "

Even the young and tech savvy are not immune; one quarter of young Canadians (wrongly) believe that Google is endorsing a source as reliable by including it in search results.

With all this incorrect information, some of it actually outright dangerous, surely it's being corrected. It is not. 

As I wrote previously, AI is not held responsible for AI mistakes; humans are. And if you're researching your health, the human is you.