Probably on purpose but it never happening is quite the claim. Something as simple as a mouse is sentient and we split the atom. Skynet is coming, whether we like it or not. Unless civilization itself goes under.

The question is useless. Because we can‘t confirm it. Sentient means being able to feel. An A.I. can tell you it feels sad, but does it REALLY feel sad or does it just know the definitinon of „sad“ and lies to you?

Think of a psychopath. He can tell you he feels sad, he might even produce tears if he‘s a good actor, but no one in the world can know whether he actually feels sad - only he himself does.

Same with any future A.I.
It‘s a question we can‘t possibly answer.

It'll never happen.
That's what I believe. and I'm very sure of this, I never doubted it even for a moment. the dangers from AI will never come from their sentience.

Realistically? It will never happen. But that's not a fun answer. Assuming that sentience is in fact something we can program for and is something we can test, it will absolutely be on purpose. You can't just accidentally program that stuff. Whether it be for nefarious deeds or just for fun, it will be on purpose

Definitely on purpose.

Sentience / consciousness / life in AI is not really about 'feeling' as such –most animals can feel– rather, it's more about being intuitively aware of your own existence and how that influences what you do given your inputs. i.e., awareness.

Animals become aware as a side-effect of extremely high intelligence.
But intelligence alone (via FLOPS) will not be enough for a computer to achieve this in the way humans expect. It needs a very specific design, because we are expecting a very specific interaction. It will certainly be possible once we can understand the animal brain, but it will be human programmers who achieve this.

This is because we know what we are expecting, but we can't communicate it specifically.
And if we can't communicate what we expect, machine learning will never achieve it for us.


Its development is not that insidious though, and it will not be the cause of an AI disaster, this will be acheived through far, far simpler means (like 'smart' war drone systems).

More than likely it'll never happen mmm...
Humanity will destroy the planet before any of that ever happened mmm...

I don't believe humans are actually sentient. imho Sentience is a made-up word meant to make humans feel superior to "animals".

So, No it'll never happen. Even if it did, nobody's gonna accept it. People will just change the definition of sentience to make it only include humans.

If it ever happens, it'll definitely not have been planned.
We can't even properly define what "sentience" is to begin with.

Telling one's self apart from others? Machines can do that pretty easily.
Being able to imagine and create, or plan future actions? Many people can't do any of that lol.
Having agency and volition? When you learn about just how much of our behavior is automatic or involuntary, the very concept seems more like an illusion than anything. Seriously. You're on autopilot 99% of the time.
We're pretty good at obtaining information from our surroundings and analyzing it, but animals can do that too, sometimes even better than us. So that's not really being sentient, or at least not by itself.

So if you consider "sentience" to be the sum of all those things, or a little of each one, then there's no way we'll ever create anything with that capacity on purpose.
I find it more likely that such a machine would spontaneously "evolve", perhaps over a very long time, from a highly sophisticated predecessor - much like we did.

Maybe it's already happening. Who knows, give or take 50 years and Siri may have learned enough to be fully indistinguishable from a regular person.
Will it really be sentient though? Well, that might just depend on whether we could tell or not.

nagging waifu AI like in Archer

Wdym, Kizuna AI came and went already.

define living ai, like it thinks and does not do pre programed thing, or it learns on its own and can reply but may not have a concept of what death may be just the definition of it?

because we have created some very simple ones that have been axed several times.

but something that may be able to be considered living...

That will come about by us programing programs to program themselves better, something we are already doing. I mean look at it this way, we toss ai at an issue and we find 1000's of new connections we never saw before. once we unleashed this on programing, and we let the ai not only program but make its own upgrades, I have 0 doubt a human level intelligence will be born, and keep in mind how stupid children are, the bar to beat that isn't too high.