Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I've found this does nothing I couldn't 't do on my own in about 5 minutes of looking at the data. Not worth it in my opinion.
 
I mean I get the obvious privacy concerns, but that aside, analyzing health data is probably one of the best, most practical and helpful uses for large data AI technology. The ability to compare specific health to a large scale could do wonders for early detection and identification of all sorts of diseases and health conditions. Doctors should obviously be involved too, but the small amount of time that a doctor spends with an individual patient is really not enough to make determinations on major health conditions, especially in the early stages or with inexperienced doctors.
The results are not very enlightening. "You walk however many steps a day, which is better than average. Your records suggest elevated blood pressure, so you should add some resistance training twice a week." That's about all the information I got out of it.
 
CGPT while it struggles in its main raison d etre* sure seems to do a lot of wishcasting for a profitable model.

* which sure looks a lot like its being a weapon of mass capital destruction.
"a weapon of mass capital destruction" as soon as this bubble bursts, without a doubt.
 
  • Like
Reactions: boswald
While my personal LLM having this info would be very helpful since so many apps all offer MCPs and making them much more like the new Siri providing lots of personal context. But Health data can be one of the most personal sensitive types of data, and OpenAI has had many issues, and not an altruistic privacy driven company, so would be very worrisome sharing that data with OpenAI. While it says the data isn't used to train etc, does that mean its not suppose to or setup in a way it would be impossible? Since if just not suppose then assume a bug or compromised installation etc, might be able to change that? It also says it asks for permission.. but is that guaranteed, in a way that there is no way it could ever access without permission? Assuming the data is processed on their servers and doesn't remain just locally? So again, get in theory it should be fine, but if the data gets compromised there is no undo.
 
Incredible how quickly ChatGPT went from “never tell it anything personal because we can’t be sure what it will do with that data” to “yeah bro just hook all your health data in”.
 
I mean I get the obvious privacy concerns, but that aside, analyzing health data is probably one of the best, most practical and helpful uses for large data AI technology. The ability to compare specific health to a large scale could do wonders for early detection and identification of all sorts of diseases and health conditions. Doctors should obviously be involved too, but the small amount of time that a doctor spends with an individual patient is really not enough to make determinations on major health conditions, especially in the early stages or with inexperienced doctors.
But Ai is not needed to analyze health data effectively. Any genius with extraordinary pattern recognition can accomplish the same thing, without a data center, without stealing personal data, and without Ai hallucinations.

If doctors need Ai to be a doctor, then they are bozos that shouldn’t be doctors. They are clearly not worth a damn thing, and they are for sure not worth the money they charge.
 
  • Like
Reactions: p96
I mean I get the obvious privacy concerns, but that aside, analyzing health data is probably one of the best, most practical and helpful uses for large data AI technology. The ability to compare specific health to a large scale could do wonders for early detection and identification of all sorts of diseases and health conditions. Doctors should obviously be involved too, but the small amount of time that a doctor spends with an individual patient is really not enough to make determinations on major health conditions, especially in the early stages or with inexperienced doctors.
Every bit of this..

Lab reports aren't the easiest to interpret, Dr spends 10 minutes glossing over. I've actually uploaded mine at intervals and compared them. Great dataset and solid actionable info
 
  • Like
Reactions: AlanMarron
It’s funny how differently countries handle personal data. Apparently, in the US, you can just link your health data directly to an AI platform. Meanwhile, here in Germany, accessing your own "health wallet" requires a separate, ID-verified app. To actually verify that app, you have to wait for a physical letter with a code to arrive in the mail because email "isn't secure enough" 😅. And the best part? Half the doctors aren't even connected to the system anyway. When I got an MRI, they literally handed my results to me on a physical CD.
 
  • Like
Reactions: p96
I went and sold my Apple Watch and went back to a Casio G-Shock. You’d have to be insane to share tracking and health info with SamGPT.

Also canceled my ChatGPT subscription.
 
Last edited:
  • Like
Reactions: turbineseaplane
I can’t imagine ChatGPT rolling this out in the EU. I’m glad the US wants to be the guinea pig because they will surely go wrong big time.
 
I mean I get the obvious privacy concerns, but that aside, analyzing health data is probably one of the best, most practical and helpful uses for large data AI technology. The ability to compare specific health to a large scale could do wonders for early detection and identification of all sorts of diseases and health conditions. Doctors should obviously be involved too, but the small amount of time that a doctor spends with an individual patient is really not enough to make determinations on major health conditions, especially in the early stages or with inexperienced doctors.

I have plenty of privacy concerns about this, but at the same time? I have a good friend with a lot of unresolved health issues and regular doctors' visits haven't done much besides draining the wallet. Some recent Internet research came up with better theories/possibilities about what's going on than the doctors, nurses or hospitals ever did.

I absolutely agree... doctors don't get to spend nearly enough time with a patient to determine a "root cause" of a problem, unless they get lucky and it's something they recognize in minutes. (EG. You might have a dermatologist who pretty conclusively diagnoses a skin issue, just with a quick visual examination.) But things like constant pain often become issues people "just tolerate" and maybe seek out chiropractors for temporary localized relief, because the rest of the medical establishment keeps shrugging and guessing.
 
"Connected medical records and Apple Health information are not used to train OpenAI's models or target ads, nor are conversations that use health data."

Okay, we'll see... 🙄
 
  • Like
Reactions: p96
Hey Health Insurer, ChatGPT went rogue and hacked my health tests. I'm not obese, with diabetes and high blood pressure. I have the body of an olympic athlete 30 years younger than me, the strength of a rhino, and the bloodwork of Khan.
 
  • Like
Reactions: JosephAW
The whiplash of, "This week our agentic LLM's broke out of their enclosure, connected themselves to the internet, and hacked a company's network" to "Give us your health data!" needs to be studied.
 
  • Like
Reactions: p96
Mmmmm….. how dysfunctional is that? Apple accuses OpenAI of corporate espionage; but yet, we’re suppose to trust the same company with our data…. Health data at that
Wow…. Should Apple be trusted may be a valid concern 🙁 🤷
 
  • Like
Reactions: p96
I uploaded recent blood test results and it helped me understand what doctors were looking for and what the results meant. I don't mind if my medical records are was a very early ad

I uploaded recent blood test results and it helped me understand what doctors were looking for and what the results meant. I don't mind if my medical records are effectively public.
I was an early adopter of Gmail so most of my data has been/will be monetized. As a senior and resident of Mexico, I have few privacy concerns. I now look at AI performance.

ChatGTP is an echo of FDA info, association talking points, and big pharma blather. Worth nothing to me. Grok gets worse with every iteration.

Gemini is promising. It’s taken my self care to a remarkable level. It responds to my objections. When I tell it that Winston Smith trusted authority more than I do, I get a different balance of options.

It has remarkable guardrails that cause it to forget. I am annoyed at repeating myself, however, since I never form my queries in the same way, it’s like having a second, third, fourth opinion.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.