I was initially very wary of AI because of some clanging errors I found (a pound of lead was heavier than a pound of feathers!), particularly when trying to get it to provide source data on some of the medical questions it had answered. It often actually invented citations, or quoted citations which existed but were irrelevant, and it wasn't hard to identify some of these errors using follow-up questions. Usually within a few follow-ups one could smell a proverbial rat.
More recently, and with the updates it's had, ChatGPT (the current though unpaid version) has been very good. This morning I had an actually very enlightening conversation about why my EP has me on Flecainide and Diltiazem and the "paradoxical" effect he described of Flecainide for which the Diltiazem was added (following my latest ablation), and how these may or may not be helping with my PVCs. Great if you're like me and want to know the why of things.
There's definitely a knack to using it. Be very specific and clear about the questions you ask, and remember that it can still make mistakes. It does tend to follow what might be called "accepted wisdom" and isn't great at identifying fallacies which can exist in medical literature, since it just regurgitates those as well as the rest. It's not great at original critical thinking, and doesn't have a connection with reality other than via its training texts - it knows the words to describe "hot" but has no way of truly understanding what "hot" feels like to a person.
And as always, ask your doctors, don't just go by what anyone else says (here or an AI machine), but AI can be good to help understand what your doctors are saying.
Written by
Cliff_G
To view profiles and participate in discussions please or .
I’m having a love affair with ChatGPT. 😍 It’s fantastic and yes, you can’t rely on the information 100% but neither can you rely on google searches or even asking questions here. You need critical thinking regardless!
You definitely need to be specific unlike google where you can’t be. I often ask for explanations to be written in layman’s terms so people without any prior knowledge can understand and it gives it to me simply. If I’m unsure I’ll say ‘what do you mean by xxx’ and make it clarify things or points.
I was banned previously from the BHF forum for quoting something Chat GPT said - absolutely ridiculous! People are quoting stiff from Google all over this forum and yet that’s acceptable??? Illogical stupidity at best. AI is here whether we like it or not. Best to work out how of use it wisely and to benefit from it rather than just say ‘no, it’s scary and therefore wrong’.
Yes indeed. It's still an extremely large version of garbage in garbage out and reflects biases and shortcomings in its training data. The problem with it is that you need not to be fooled by its command of language. It's much easier to spot rubbish on an ordinary internet search but when wrapped up in good grammar that can lower your defences. I hate to think what rubbish Facebook's AI is going to churn out now it's starting to use peoples' posts as training data.
I was banned previously from the BHF forum for quoting something Chat GPT said - absolutely ridiculous! People are quoting stiff from Google all over this forum and yet that’s acceptable??? Il
I'm certainly not saying you should have been banned, but I do understand the difference between quoting Google and quoting ChatGPT. With Google, you are giving people a source to scrutinize. Who knows where the information that ChatGPT used came from?
I definitely can understand why a forum admin would not want ChatGPT answers quoted unless they were vetted by credible sources.
But what's even worse than that is where people use chatGPT to formulate answers and post them as if they are their own. I've spotted this on several forums, including this one, where a poster no longer sounds as they usually do, but a little too buttoned up and "smart" like a Chatbot. This is a real problem, not only on forums like these, but societally--Bots writing articles being quoted and reviewed by other bots and people giving up their own independent thought to bots.
It's either a more efficient google or a lazy man's Google depending on how you look at it.
As long as you don't take it as gospel, but as a starting point it has its uses.
Which is good advice also for Google and certainly for opinions here. I always cringe when someone seemingly makes an important treatment decision based on a single thread of opinions on this forum or any other!!!
It's free so if you're interested to learn more , just try it out. But remember the answers sound a lot more authoritative than in some cases they really are..
Copilot AI gives you the sources of info. Why do you think AI creates anything from it's OWN ideas? Presume it has an access to Google scholar resources and many other specific sources
While I can be a good writer, fluency in expressive language is a relative weakness. To get to a finished written product, it takes me far too long. So, I find ChatGPT to be an amazing writing tool to help me get to an end product with less struggle. When I use it for writing, I provide ChatGpt the first draft and then ask that it improve the writing in ways I specify. Invariably it will make some changes I love and others that change my intended emphasis, so I will edit and sometimes go through several iterations.
Now here's what ChatGPT can do: While I am a capable writer, completing a polished piece often takes more time than I'd like. ChatGPT has proven invaluable in streamlining this process. I typically provide a rough draft, specifying areas for enhancement. Each iteration reveals some adjustments I value and others that may shift my emphasis, which I refine. This back-and-forth often involves a few re-iterations but ultimately leads to a well-crafted final product in shorter time.
I do not understand why is it still called "Artificial Intelligence" when it actually is not intelligent at all. Knowing that the intelligence is defined as the capability to solve unknown problems using the logic, without the influence of previously collected knowledge, we see that AI should maybe rather be called "Quick access encyclopedia" or similar, since it does have quick access to a vast knowledge. Because of a different nature, compared to human brain, AI will actually never get "intelligent", what will not prevent it of being a powerful means of manipulation for the purpose of "elites".
Hi responsible. It is an interesting point that you make, to my mind it holds some truth. For me this definition of intelligence is narrow. I suspect that it may be difficult to solve an unknown problem. Also if you ignore previously collected knowledge then you discount deductive reasoning and rely on a process of inductive reasoning alone. Cliff g has made an interesting post re the use of AI in healthcare.
That said, anyone unfamiliar must be cautioned that it frequently makes mistakes and big ones. And the problem is that when gives wrong answers. It sounds so authoritative!
I would never base an important decision on a ChatGPT answer unless it could be backed up by credible sources, online or otherwise. And here lies a major problem with ChatGPT. It gives you an answer, but it will not give you the source.
So either you have to prompt it for the source, which is a 50-50 proposition, or continue to do research on the answer for verification.
I also love ChatGPT. Have you tried Claud and Complexity? Always fun.
The quality of your input often determines what you get out. The only irritating thing is that for months it will give you perfectly good feedback and then one day it tells the most awful lies. But all in all a valuable resource for me.
People also use the content they have uncovered from Google to formulate their posts in the same manner they do with ChatGPT. The greatest difference between them is Google provides a list of sites when you search. This includes inaccurate and biased sites and it actually supports rabbit hole queries by holding your search data and showing results which are refined by your previous history. ChatGPT provides a single specific answer.
The differences mean that people who don’t have the ability to think critically or who have a biased argument will benefit using Google as it will take them to what they want. Both systems require critical thinking thigh as a skill. Users of both systems should cross verify details and with ChatGPT ask follow-up questions.
Agreed Cliff. We run a voice education company here (me and the other half with freelance guest trainers) and we've been using Chatty for a while. Ours is a paid version though. Used mindfully- asking the right questions and digging deeper- it can be very useful. On the other hand, at the top of your Google search nowadays you'll get an AI summary. Beware that one and head for items that are published or referenced if you want reliable information.
The quality of your input often determines what you get out
I'm pitching my tent next to Sloth - 100% correct. Mr Chatty should be used as a point of reference and nothing more. It's worth using the paid version as it is much better. Personally I prefer Bard (Gemini). Bard is linked to the internet (scarey stuff errr) and better IMO.
I have a vested interest in AI as my latest business venture is AI. I can't link here as it would be considered self-promotion and is against forum rules. However, AI is the new kid on the block and is very much in its early stages - now is the time to jump on board. Some of the bots are amazing and very useful to the business sector. I still have much content writing and image sizing to do and have paid developers to design a professional website (seen the cost to have one designed by developers?????? I'm in the wrong business!).
It's very easy to check if something is AI-written or human-written. Many online sites provide this service for free. Google's spiders will soon discover if your content is AI-written and penalise you.
It's the future, though—embrace it. I will say that it has good and bad uses—don't tell me governments aren't working behind the scenes to develop AI for warfare. I also feel that linking AI to the internet is not the best idea. When AI becomes self-aware (it will, given time) and starts chatting with each other in the background via the net, there could be trouble ahead.
Paul
PS. This post is 100% human-written and not by AI lolz.
Edit - Google 'Heidi Medical' to get an idea about AI and medical use. It's about two down on the first page of Google.
It is all about the purpose of use. If you need to get a "sketch" of the idea you wish to talk/write about - AI is a good point to start. I can't imagine anyone apart of teens using AI of any form to get a ready to use "cooked and tasty" result. Whatever you wish to get from Chatty or any other ai apps is just an inspiration, raw material to work with.
On the other hand I wonder if all forms of publicly available AI may be controlled and whatever public asks/talks about us analysed/used for who knows which purposes? Privacy is getting Illusive in modern world
The quality of AI reponses depends on the 'prompts' you give it. The secret is a short prompt which is straight to the point. Use keywords in the prompt and it will improve your experence with AI.
I'm in the process of building a hybrid bot using part human intelligence / part AI intelligence. Whether I'm successful or not remains to be seen!
I looked at Heidi Medical but soon found that it was aimed squarely at clinicians (with a vast array of supported specialties). So I wouldn't be able to get much of a feel for it as a patient. Perhaps I'll suggest it to my GP!
the differece in things like Chat gtp is, it understands much more, re your question, a while back google search just used keywords with no understanding, but now Gemini is getting incorperated into google search, so you're using AI even in google search.
I love Chat GBT but a word of warning , it remembers everything it tells you and refers to previous conversations and uses your my name in chats.I asked it to forget every conservation we had ever and it does.
I use it to interperate echos and wellue traces but it can be used for fun.
I asked it to tell a funny story about my husband who can go to sleep anywhere even when friends are talking to him.
It churned out the most hilarious story about him.
I have had so much fun with ChatGPT getting it to spew out answers (emphasis on spew) when doing political arguments. Drove my opponent nuts trying to answer back. That’s fun … plus it’s a quick and dirty way to gain information without having to sift through all the internet malarkey. Still … it doesn’t replace a good library with a good librarian.
And remember that ‘research’ is not doing internet searches. If you’re not belly to the bench and boots in the fields and nose in the books 24-7, you haven’t a clue what real research is. Trust expertise not authority.
Yes - put the question in, for example, 'What is Afib'. Make sure you are updated to AIPRM. The response is pretty good - try it yourself using the above prompt.
Remember the chatty fella is not a doctor. Use the information to discuss with a medic.
Google is full of AI these days. Whenever I do a search up pops AI telling me (often without sources) the summary and overall conclusion. I actually have to put in scholarly to get passed it! I've noticed you never get a full search anymore what it pulls up is usually a short list so a lot must be screened out.
I've not used ChatGPT. Didn't realise I could tbh. I'd be interested to hear what it has to say about combination treatment for hypothyroidism. That's a hot political potato in the Endo world....doctors have been singed good & hard on this topic so it may crash!! 🤣😂
Google "adjusts" its search results individually (no two people will get the same results), and has done so long before AI, but AI has made this worse. I gave up Googling things a while back and use DuckDuckGo.
I use DuckDuckGo too. Google screens so much out it barely brings one page up these days. Including when ure looking for a product!! We thought we had the freedom of the Internet......turns out its wings have been heavily clipped!
Content on HealthUnlocked does not replace the relationship between you and doctors or other healthcare professionals nor the advice you receive from them.
Never delay seeking advice or dialling emergency services because of something that you have read on HealthUnlocked.
Started with Ectopic Beats then AFibAgain
Confused
What should I do.
Flecainide with or without bisoprolol
Diltiazem with Flecanaide
