They should ask the computer what it thinks about self-pleasuring. But they won’t.
My idea of self-pleasuring is a hot cup of coffee and a nap.
1) A tremendous amount of this crap is NOT AI.
2) AI has a specific definition and has for about 40 years. The code must learn something from external inputs and THEN REWRITE ITSELF. The code must change. So say goodbye to configuration management, but that’s the definition. The code changes itself.
3) Most of what is presented as AI is just fast software. You could do this stuff, a bit slower, 30 years ago and no one would have called it AI.
4) The ChatGPT stuff is largely not AI. It is a hyper fast searcher of the internet, and this is a valuable thing to have, but it is not rewriting its own code. The only learning induced behavior I have seen is something like this:
How many people suicide on the London subway tracks?
It then answers, if it can.
Then . . . How many in Boston?
It will then understand that this is from the same IP address and refers to the previous question. So even though suicide is not mentioned in the 2nd question, it learns from the first and understands suicide is the question. That’s not AI. That is hyper fast searching.
5) Understand the value. Hyper fast searching. This claptrap about asking it about the existence of God or are Democrats moral or all that kind of thing is stupid.
Understand what is valuable. If you would like to know the average ferryboat travel time between Juneau, Alaska and wherever, you can spend an hour searching, or you can ask ChatGPT. If you want to know if deer population in northwest Wisconsin is greater than last decade, it can find that information quickly.
It can make errors, but if you ask the proper questions, they will be few.