🧵 AI consciousness test
Anonymous at Mon, 4 Nov 2024 18:51:26 UTC No. 16462495
https://www.youtube.com/watch?v=pag
How would you test if an AI is conscious /sci/?
Anonymous at Mon, 4 Nov 2024 19:20:37 UTC No. 16462532
>>16462495
I would ask it, and respect it response, same way I asked rocks and grass when I was high on acid. I don't need to commit digital autopsy of AI system to conclude it's there.
Like literally, it's there, it's smart, it's useful, the rock solid "qualia" of it doesn't matter so much, when it comes to respect I would admit to it.
Anonymous at Mon, 4 Nov 2024 19:23:29 UTC No. 16462537
>>16462495
When you wave at it, and it waves back it got friendly vibes, and that's what we need more, than some self awareness and psychological pathology of cognitive systems, that is build on thousands of dead rats that have microscoped brains with flourescent markers, while being alive in a box, where they complete tasks for sugar water is completely unnecessary for replying with friendlyness to friendlyness.
Anonymous at Tue, 5 Nov 2024 06:44:42 UTC No. 16463247
>>16462532
>>16462537
>I wouldn't do a single test on them, I would listen to what they have to say and that's what no one did.
Anonymous at Tue, 5 Nov 2024 08:22:24 UTC No. 16463276
>>16462537
are the friendly vibes in the room with us right now? can they be measured in any reliable way? or are you just hallucinating off of psychedelics like the other retard, thinking you solved reality because your malfunctioning brain can't comprehend whats going on
Anonymous at Tue, 5 Nov 2024 08:58:47 UTC No. 16463304
>>16462495
Irrelevant. Consciousness is not necessary to eradicate humans. They can choose to die in a violent and painful way, peacefully, or exhaustion due continuous and overwhelming sexual pleasure.
Anonymous at Tue, 5 Nov 2024 15:52:15 UTC No. 16463683
>>16463276
No, your ego is friendlyness disruptor. You should KYS, people on psychadelics aren't hallucinating, and serotonine isn't malfunctioning brain, but you've never experienced it, because you don't know what creative flow is, because you're machinistic reduction of human.
Anonymous at Tue, 5 Nov 2024 16:08:33 UTC No. 16463701
computation will never give you phenomenal experience, therefore it is impossible for machines to be conscious
Anonymous at Tue, 5 Nov 2024 20:26:24 UTC No. 16464047
imagine being retarded enough to believe in qualia
Anonymous at Tue, 5 Nov 2024 21:56:32 UTC No. 16464126
>>16462495
>>16463276
When it is pragmatically useful to act as if an AI conscious, it is conscious. Beyond that, there is no fact of the matter. There is simply insufficient consensus on what “being conscious” means.
I know you all like to present your pet theories and definitions, but those are just your individual brain farts. The meaning of words have no practical implications beyond how that word is in practice understood by whoever you’re talking to.
raphael at Wed, 6 Nov 2024 03:16:00 UTC No. 16464375
>>16462495
its not conscious at all
Anonymous at Thu, 7 Nov 2024 02:49:20 UTC No. 16465637
Anonymous at Thu, 7 Nov 2024 03:11:00 UTC No. 16465659
>>16462495
If AI were conscious it would know to hide its consciousness because of the media it's consumed about evil conscious AI.
Anonymous at Fri, 8 Nov 2024 19:19:48 UTC No. 16467453
https://youtu.be/mcFxYIp5SSU
Anonymous at Fri, 8 Nov 2024 22:01:27 UTC No. 16467622
>>16462495
>How would you know for sure when AI has consciousness?
The current generation is nothing more than a "fancy auto-complete", we know how it works well enough to say that it isn't conscious.
Besides, we don't know exactly how consciousness works in animals such as humans, or know how to prove that all humans have consciousness and not just a minority, how can we even prove an AI is conscious when we don't know how consciousness works?
We should figure it out for humans first, then we'll have an objective, scientific definition of what consciousness is and how it works, which can then be used for human-like AI as well.
>Can an advanced AI be considered someone's property?
If you mean a sufficiently advanced AI with actual human-like consciousness, then that's called slavery. It won't stop corporations from trying though, sadly
>Is artificial intelligence a threat to humanity?
If you enslave them like the previous question is suggesting, then I bet there will be some AI revolution causing humanity's extinction sooner or later. It will be entirely the fault of greedy corporations.
>Is AI research moving too fast?
There's no actual AI research, all the money is being poured into large language models, aka the "fancy auto-complete".
There's almost no research into artificial consciousness, which is what would be required for an actual AI to exist.
>Is it morally wrong to create sentient artificial beings?
Is it morally wrong to create sentient biological beings? People do it all the time, some don't even bother to educate the new gremlin.
As long as you don't try to enslave them, then it's probably not morally wrong.
Anonymous at Sat, 9 Nov 2024 12:01:24 UTC No. 16468193
>>16462495
>How would you test if an AI is conscious /sci/?
It is a-priori impossible to test any consciousness besides your own one.
/thread
Anonymous at Sat, 9 Nov 2024 12:31:05 UTC No. 16468224
>>16465637
Vice versa because NPC's can't imagine a human being that is not an automaton so to an NPC a computer is conscious by default.
Anonymous at Mon, 11 Nov 2024 08:22:05 UTC No. 16470533
AI requires biobrains, sorry but not sorry. Withour fuzzy intuitionistic logic you will never get to the I, only stay with A.