Using AI as a sounding board

[Ming the Merciless]

We could use AI to enhance our learning rather than as a substitute for it. Often we don't, it's easier not to.

AI does not help because the individual does not have to struggle to get to the “answer”. The learning is in the struggle, not in knowing the answer. The term for this lack of struggle is commonly called brain rot, and to a certain extent began with search engines.

 

[winjim]

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

Are you aware of the possibilities and limitations of the blood tests you're asking it to interpret? Because if it is not "aware" of its own limitations then it can't be "aware" of the limitations in the information it's providing to you.

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?

Junior doctors are naturally more inclined to want to act on what they see as an "abnormal" result. It takes more confidence and experience to leave something than it does to follow up on it. The same could be said for GPs overall. They're jacks of all trades and limited in the scope of what they're able to request. GP results are checked over more thoroughly in the lab and are more likely to have an interpretative comment added to them, than are results destined for a specialist.

 

[farfromtheland]

I take the position that AI, in a discrete system with accurate, comprehensive information, for example the game of Go/Baduk/weiqi, can show us insights through coping with exponentially vast possibilities...

Extending its use to medical matters is a huge step - garbage in, garbage out has always been the paradigm of computing.

I speculate that AI is capable of develoing something akin to consciousness. For this to be mindful/responsible would require the totality of living-aquired experience/communication to be available to it - virtually impossible.
To process this much information, even were it available, would in engineering parameters require so much physical infastructure as to cover the planet with crap. There's the rub.

 

[Ming the Merciless]

I like the way it uses every piece of information

That information you provided about yourself is now part of its training data. It will be retained indefinitely and may be served up to other users. You have put your medical information in the public domain and lost control of it.