Do not be fooled by AI doctors

“I Googled it!“

“I googled it” is a sentence that probably every doctor hates to hear. And probably all of us can admit doing. At the same time, we know it’s not always the most brilliant move. “Doing your own research” on medical conditions by writing some symptoms to Google search box results in various illnesses. The results often diagnose us with the most critical diseases – cancer, heart attack, HIV, and meningitis. Because, of course, sore throat and fever are the most likely symptoms of HIV. Or maybe not.

Now that AI hype is everywhere, the media has started writing news articles such as “How a mother used ChatGPT to diagnose her son’s illness” and how “ChatGPT can replicate your Doctor surprisingly well”. After seeing these headlines, it may look like we should all use ChatGPT to make our own diagnoses.

Before following that path further, I suggest you hold your stethoscope for a moment.

A chatbot can work only with the data shown and a diagnosis based on what you tell it. The chatbot has no capabilities to check you physically or give you a proper medical opinion. It doesn’t differ from Google searches. At least not yet.

But let’s continue the train of thought a bit further. What happens once AI does work properly? A couple of scenarios come to my mind.

What if the AI bot is free to use, but behind it is a large corporation that wants to sell its drug products? The laws allow you to sell over-the-counter products at supermarkets. Why not have an AI bot that talks with you about your medical problems or worries and suggests the latest Doctor John’s medical “iodine+calcium” tablets – “These have the chance of helping you with your condition.” Sure, there might be a chance, even though the likelihood might be near zero. It is not technically lying, but questionable. And this is already possible to do.

How about taking it even further? We already know that there are ethical issues related to the drugs the doctors prescribe to us. The doctors might be incentivized to prescribe a particular drug company’s brand instead of a cheaper one. Perhaps the doctor gets a bonus or commission for each prescription of that specific drug.

What happens when that big pharma company starts to offer their properly regulated, Medical Device Certified AI Doctor for free? What if that particular AI doctor can prescribe drugs for common illnesses? And who cares if the prescription is for the most expensive brand that just happens to offer the free aI doctor? At least the doctor’s appointment was fast and free.

And if we continue further down the rabbit hole, we come to even stranger situations, especially regarding treatments. The internet is already filled with various alternative therapies (Unicorn Healing, anyone?!). These cheat people into doing and believing the weirdest things. This is a real possibility and doesn’t differ from fake news. Imagine having an “AI alternative medicine Doctor” or a “Regular AI doctor” created to spread someone’s questionable agenda. An AI shaman that tells you to drink silver water or to use magical stones to get rid of fever. There sure is money to be made here. And quite sadly, it is quite likely there are people willing to do these scams.

Yes. This is the thing I am worried about. Every news article about using AI to self-diagnose or replace your regular doctor also gives room for the discussion of finding your alternative AI doctor. Soon, there will be good AI doctors with proper medical device certifications, but many also many that aren’t. Even those accepted as medical devices can have biased agendas to direct us to particular products or other actions that we might not have done otherwise.