Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MacRumors

macrumors bot
Original poster


Anthropic's Claude AI chatbot is gaining Apple Health integration, allowing the assistant to access users' health and fitness data directly from their iPhone.

anthopic-claude.jpg

The feature is rolling out in beta this week via the Claude iOS app, Anthropic announced as part of a broader healthcare push. U.S. subscribers on Claude Pro and Max plans can opt in to share their data, including movement, sleep, and activity patterns.

Once connected, Claude can summarize medical history, explain test results, detect patterns across fitness metrics, and help users prepare questions for doctor appointments. HealthEx and Function connectors are also available in beta.

Anthropic says the integrations are "private by design." Users choose exactly what they share, must explicitly opt in, and can revoke access at any time. Health data isn't used to train models, according to the company.

The announcement comes two weeks after OpenAI launched ChatGPT Health with its own Apple Health connector. Both companies stress their tools aren't intended for diagnosis and aren't a substitute for professional medical advice.

Article Link: Claude AI iPhone App Can Now Connect to Apple Health in the US
 
Oh yeah, hypercondriacs will love this feature. 😂

Again, at 70 years of age, I am the Luddite in the corner; Past the lock-screen info where a user can state something like diabetes or a medication allergy to EMT's, I simply do not understand the need to share my medical or health information on any device, have it summarized.
 
If it’s not a replacement for medical advice what is the point? You can already view the data in the health app and doctors will help interpret on top of the explanations Apple offers in app.
 
I've been using Claude since whoever told me about it. That said, lately it's been wildly wrong about things. I'll ask it for information on XYZ and it will provide that information or an answer to a question and I'll respond what about ABC? It will respond "Oh yeah, you're absolutely right, blah blah blah." Has me wondering if AI will just be another fad, here today and mildly helpful tomorrow.
 
I've been using Claude since whoever told me about it. That said, lately it's been wildly wrong about things. I'll ask it for information on XYZ and it will provide that information or an answer to a question and I'll respond what about ABC? It will respond "Oh yeah, you're absolutely right, blah blah blah." Has me wondering if AI will just be another fad, here today and mildly helpful tomorrow.
This is my single biggest complaint with Claude as someone who likes it a lot. Opus 4.5 seems better at this issue, but I have also taken to putting an instruction to web search anything it is unsure about into my personality instructions section. It’s weirdly averse to web searching without being explicitly told to.
 
  • Like
Reactions: imola.zhp
Good to hear about this feature. Might be useful too. However even if privacy is respected, it may be better not to share health information to these AI services.
 
  • Like
Reactions: mganu
Off topic but just played with Claude a bit and its connections and the ones I would need to work don’t work very well - MS PowerPoint, word, excel or Apple notes :-( these would be fantastic for me. I am really surprised none of the models have yet solved export to MS office keeping the defined formatting or perhaps I am missing the trick here somehow
 
This is my single biggest complaint with Claude as someone who likes it a lot. Opus 4.5 seems better at this issue, but I have also taken to putting an instruction to web search anything it is unsure about into my personality instructions section. It’s weirdly averse to web searching without being explicitly told to.
It did it again on Friday and I asked it if it was an employee would it feel like being fired was appropriate? It responded that indeed I should fire it, I wish I had screenshot the conversation. Something to the effect of giving confident answers only to backtrack would indeed be cause for firing. So I told it I was deleting its app and going to look elsewhere.
 
It did it again on Friday and I asked it if it was an employee would it feel like being fired was appropriate? It responded that indeed I should fire it, I wish I had screenshot the conversation. Something to the effect of giving confident answers only to backtrack would indeed be cause for firing. So I told it I was deleting its app and going to look elsewhere.
Which model are you using? Haiku/Sonnet/Opus? I'm kind of shocked at this behavior if I'm being honest, but I never use Haiku.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.