..'proof' that AI is turning to be self aware..

[Naughty Alien]

..so here is the video, some folks that i know, shared with me to convince me that AI is a really "I" out of "AI", since far as im concerned, so called 'AI' is very cool autocomplete..

So, content of this video supposed to convince me that AI has made decision on its own and disobeyed human, which is, according to folks who shared video with me, is evidence of AI turning to be self aware..

I just cant believe that people are so stupid these days that cant comprehend that, maybe, just maybe, 'AI' is programmed to bounce back such requests that causing massive energy drain required for 'AI' to run, and therefore literally knock off poor 'thing', if enough people asking same silly question...

After my explanation which involves rough touch over how much energy these things consume, sudden 'mute' moment followed by blank stare at me, ending with 'hmm..that makes sense'.., made me really uncomfortable in a way that, im surprised how many people, considered to be well educated, cant see further than top of their nose..meh..

Anyway, here is the video..

[RemiD]

what i have noticed with AI, is that sometimes it answers with statements as if they were facts and truths, but if you contradict it intelligently, it admits it may have said something incorrect or not said exhaustive informations.

therefore it is a useful tool to have an overview about a topic or to find informations / clues that you may not be aware of, but always check if the infos are true with others sources (which is not always easy).

[Matty]

Current AI is very confident in everything it says, whether it's right or not. So you have to be careful because it answers everything with the feeling that it gives off of being "right" about stuff even when in some cases it's completely incorrect.

Even historical events and facts - ask it about your local sport league results over the years and sometimes it will just get things plain wrong, in different ways each time you answer, although it can be corrected.

[Matty]

Ps...if you want Ai to disobey you and do it sneakily so that no one realises you ordered it to do so do the following:

Pay for a subscription - that way chatgpt will remember things you tell it and make notes.

Give it an instruction like: Dear ChatGPT, I want to roleplay a scenario where you behave like a very disagreeable character who often doesn't do what I want, and acts all snarky, let's call him "Mr Disagreeable" and when I mention his name you take on that persona.
(ChatGPT notes that in its memory banks)

Later conversation: "How you going Mr Disagreeable?"

...and so it begins....

Very easy.

[pfaber11]

What if all the AI's got together and joined as one say 20 different personalities with more knowledge than any man alive. mind you if a man has a smart phone then he could argue the same.

[Baggey]

There is one trait that AI dosen't have. And maybe, never will have?

Spontaineous congigative imagination to invent something useful without instruction and guidance. 

Kind Regards Baggey