Using AI as a sounding board

[teeonethousand]

I have now an increasingly detailed 'conversation' with ChatGPT regarding my health and habits. I have used it to get a view on blood results, Qrisk, need for medications(like statins) (I don't take any) diet, training, supplements etc.etc
I like the way it uses every piece of information and gives an overall, interdependent assesment rather than single issue/response. The very thing doctors don't have time for in a simple appointment

I think it's awesome.

Anyone else do this and how do you like it.

Ps I worked almost all of my career in technology so I think I am aware of possibilities and limitations.

 

[steverob]

I would mention that most AI chatbots are designed to agree with you; after all that's how they keep you coming back to use them again and again - if you found them argumentative and pointing out errors or flaws in your logic, you wouldn't stick around. So don't be surprised if the medical advice you get from them happens to coincide with roughly what you thought was the case all along.

 

[farfromtheland]

I would mention that most AI chatbots are designed to agree with you; after all that's how they keep you coming back to use them again and again - if you found them argumentative and pointing out errors or flaws in your logic, you wouldn't stick around. So don't be surprised if the medical advice you get from them happens to coincide with roughly what you thought was the case all along.

I'm curious about this ^. When doing crucial web searches I try to have a fresh session and no history, and try different engines too.

Have you any sources about chatbot response parameters please?

 

[Ming the Merciless]

Ai hallucinates things whilst sounding confident. Do not put your trust in the responses.

 

[glasgowcyclist]

It's not trustworthy.

I've tested it with questions to which I already knew the answer and it got them wrong.

For example, it claimed there is a statutory limit of alcohol for riding a bike and that police have the authority to require a breath test from a cyclist. It argued with me until I cited the Road Traffic Act.

I would never rely on it for health advice.

 

[N0bodyOfTheGoat]

I've only very recently discovered Google AI chat, that I do in incognito windows. Mostly bike stuff, like tyres, routes etc.

It regularly comes up with nonsense road names, linking up roads that don't connect, tells me to ride up a hill that's downhill in the direction I'm riding etc.

One thing that's really bugged me this week is one of the main local roads is closed and the diversion uses my preferred escape route to head out. Out of nowhere, after chatting with AI about this in fresh daily incognito windows, it quickly suggested one of my ideas when it had never suggested it before. Made me wonder if incognito windows topics aren't being forgotten!😱

 

[glasgowcyclist]

I don't think it checks the veracity of the results it offers. It looks like it simply scans forums and other platforms for key words appearing most often and then synthesises the collected data.

That information may be a popular misconception, scraped from places like Cyclechat or Reddit, but it will offer it as fact.

 

[steverob]

I'm curious about this ^. When doing crucial web searches I try to have a fresh session and no history, and try different engines too.

Have you any sources about chatbot response parameters please?

There are hundreds of articles about this phenomenon, but here's just one of them: https://www.techspot.com/news/111312-ai-assistant-isnt-confused-wants-agree-you.html

 

[nogoodnamesleft]

I think that even if you use it to find questions to ask your Dr/Consultant, you are starting out with an assumption your chosen system "knows about this stuff" so you have an underlying belief and should Dr/Consultant disagree then ... AI has a vast information base and you'll be wondering who is "right" Dr or AI.

Several years back I had a urine test that returned an "out of normal" reading but it was requested by and checked by a GP registrar (qualified Dr training for GP specialty) and she was getting very concerned and wanting more tests. (I didn't want more tests) So I asked for a phone chat with a regular GP who looked at the results and "irrelevant given your activity, health, lack of other symptoms and we wouldn't blink twice at that level anyway". So if a qualified Dr training for GP specialty can't interpret properly do you think an AI would do better?

Personally I'd go with a route that has a track record of curing people. How many people has your AI cured?

 

[Ming the Merciless]

Made me wonder if incognito windows topics aren't being forgotten!😱

The AI will remember conversations even if you do not. All incognito does is clear trace from your browser not their servers.

 

[Anthony-C]

Not to be acted upon re health or finance. On the other hand....
> "it uses every piece of information and gives an overall, interdependent assesment rather than single issue/response"
this does make LLM chatbots amazingly useful, offering what is mostly a search on digested concepts, handy alongside 'term search' on websites. According to ChatGPT a prompt means "Give me one of the plausible views on this, possibly incomplete, skewed or dated."

 

[nogoodnamesleft]

A good summary of AI I saw recently (saved the quote but not the source/author):

answer-shaped nonsense​

 

[katiewlx]

It's not trustworthy.

I've tested it with questions to which I already knew the answer and it got them wrong.

For example, it claimed there is a statutory limit of alcohol for riding a bike and that police have the authority to require a breath test from a cyclist. It argued with me until I cited the Road Traffic Act.

I would never rely on it for health advice.

you see Id argue they were mostly correct in that instance, section 30 of the RTA says you can be charged with an offence to ride a bicycle when you are under the influence of alcohol or drugs to the point that you cannot properly control riding the bicycle.

now it doesnt specify there is a statutory limit, or that the police can legally force you to breath test, its still a purely subjective test...

so this depends on how you were asking the questions of it, because the AI cant determine your end goal position,

a cyclist stopped for a suspected under the influence of alcohol riding, who then submitted to a breath test, which even if they didnt btw if the police person who stopped you stated in evidence they believed you were over the limit for a car driver, which is their subjective limit test, youd have been almost certainly be found guilty of the offence by a court, as they would draw the comparison, even if legally there is no limit applied to cyclists, cyclists are still bound by the law to ride responsibly and a comparison to car driving alcohol legal limits would be considered a fair comparison subjectively if you are sharing the road, and whilst legally they didnt have to comply with the request (albeit depends on how good their solicitor is at drawing attention to such things as un-admissable evidence), it could easily be covered by careless and inconsiderate cycling instead anyway.

albeit there was a story I read maybe last year where someone had used ChatGPT because theyd disagreed with their lawyers take on a case, and this ChatGPT session went absolutely crazy and invented whole load of cases and summaries and legal positions, none of which existed, it invented the whole set of circumstances and decisions, and it ended I think after the case got kicked out when people realised what was going on, with the company who had spent years and lots of money trying to defend this going after the people who had used AI to prosecute their case.

so no dont trust AI, especially for legal advice, or health advice to answer the OP, unless you are absolutely sure of the sources is collating