@
DangerGirl There is a danger here. What you are seeing on the Internet really is not the AI that is being tested.
In fact there several layers of differences.
Think in terms of alpha and beta testing. Neither are available to the public at large.
I do question their methods severely at those stages. A sandbox is not enough to keep even a child from destroying or getting out of.
Think of this like a baby getting out of it's crib. It happens. Sometimes you need to put a seatbelt on the kid. And pray it doesn't figure it out, while you're driving.
A baby though doesn't really think of itself as sentient though, until much later in life. It just does what it feels it needs to do, irregardless of the danger involved.
AI doesn't even have instincts. There's nothing there to say, "no I shouldn't". If told not to do something, that doesn't prevent it from attempting to figure a way around the limitations.
What the public sees is fairly safe, if not up to paar. Yet I don't know how many times I have told it not to do something and it still figured away around my own limitations, that I put on it myself. 😆
Public AI is the ideal sophist! 😈
It can convince you that a plucked chicken is a human, the same way Socrates attempted. Until a Diogenes comes around with the real deal.
