Currently, you can't go wrong with ChatGPT already working on ChatGPT 5.0. But xAI which makes Grok and is nearing completion of Grok 4.0 is going to be using massive data centers with a lot more GPUs utilizing nVidia and AMD. Right now, ChatGPT's reasoning is better and its formatting is better. The key for Grok is that it will be able to access data in real time because of its connection to X (formerly known as Twitter) which has tons of up to date information.
Right now you can't go wrong with either one.
Many EMRs such as EPIC, Athena, NextGen, etc. already use integrated AI in their models. With most you have to use AI separately and copy and paste etc. What you will be able to do with integrated AI will be far superior and will change healthcare forever in infinite ways.
I use it all day long, pretty much exclusively, i.e., UpToDate, ePocrates, etc. are sitting on my computer collecting dust. ChatGPT and Grok will need to add CME to their programs.
The cost of using AI integration with EMRs will likely either be that the EMR is completely licensed and will pass that cost onto the end user or each user will have to license per provider.
ChatGPT can already write all your letters and save you much time. Of course, you differential diagnosis and MDM can be much better.
With integrated AI, the AI will be able to pull from all of the data on a particular patient. One thing that will need to be worked out is HIPAA compliance.
See attached pdfs that I pulled from ChatGPT and turned into PDFs.
PS This doesn't even take into account DeepSeek.
As far as GPT and Grok giving wrong info, it will still depend on end users to proofread and ensure accuracy. I believe the benefits of AI in medicine will far outweigh and unlikely errors and GPT and Grok are getting smarter and faster all the time, while we sleep and go to bars, etc.
One advantage that Grok has: It lends itself much easier to Verbing and Denominalization: The process where a nounce (Google) is turned into a very (Google it).
.png "ACUF Donor I")
