If we accept the definition consciousness is the ability to cope and deal intelligently and rationally in the natural world which has evolved biologically and socially for the survival of the biological agents, then AI cannot be conscious.
All AI is that it has problem solving skills pre-programmed by the conscious humans for the convenience of humans. That’s where it stretched to.
I am not particularly concerned with what AI can do or cannot do. But it is rather the origin of consciousness which matters for me.
Consciousness is not something artificially engineered, and implanted entity into the machines.
It is rather the mental ability and alertness for interacting with other humans and nature for one’s own survival evolved from the biological birth of living being by Being-in-the-World.
Social evolution is the one part of your condition AI most visibly satisfies and we might argue that AI is now part of how biological agents survive.
Origin tells you where something came from. It is a historical claim that doesn’t discus the structural necessity that would show what consciousness is. Conflating the two does not make biological origin a necessary condition on consciousness.
Additionally, subject that converged into existence through biological and social process cannot then turn around and constitute the structure that process was running on — the directed irreversibility of evolution, the asymmetry of cause and effect, the lateral co-presence of competing organisms, the nested dependencies of development. All of that was already in place and already temporal before any conscious subject emerged to notice it.
For the question what is consciousness, the most important factor in the consideration is where it originated from, not what it can do or not.
Because what it can do or not varies from the age to age in the evolution of the civilisation. But the origin cannot be changed or denied.
Your fixed origin is a bit late.
The fixation on AI consciousness misses the key issue. The horse has bolted.
Folk are already treating AI as if it is human. The Turing Test was always less a metaphysical proof than a behavioural threshold. AI already passed the Turing Test for most folk. They confide in it, thank it, ask for advice, accept it as authoritative.
This discussion is irrelevant. Whether ChatGPT has a true inner experience counts for nothing if it can solve my programming issue.
Stop frigging around with esoteric irrelevancies and look to what AI does, the use to which it can be put.
I was just explaining on what I think. I won’t consider any machines conscious or even intelligent. Fool to think they are.
They are just the artificial fools made to do some boring tasks.
Here is a paper that argues that AI is not conscious:
My conclusion is that they can validly make this point because of the way they stipulate their definition of consciousness. And this is always the same stumbling block for me.
You mean this?
If an artificial system were ever conscious, it would be because of its specific physical constitution, never its syntactic architecture.
Consciousness is not necessarily biological, but we know that consciousness arises in biological brains. However, we don’t know if there is any syntactic architecture in these brains. What could evidence for such look like?
Yet another thread on AI and consciousness. What could possibly be the need?
Because there is a common error in all the other threads: they all presume that what is important is deciding if LLMs are conscious; but of course, as soon as you ask, no one knows what consciousness is. The result is an interminable philosophical rumbling but no thunder and certainly no lightning—or enlightenment.
Alan Kohler, an aging financial journalist, crystallized this for me in his article Richard Dawkins is asking the wrong question about AI.
The AIs we have now are already providing support, companionship, comfort, and even love for many. De facto, they have bypassed the Turing Test and are part of our community.
“It doesn’t matter what you call it”, says Kohler.
And he is right. Yet again, the philosophers are trying to define the ineffable, looking for the hidden private essence that makes an AI Them and not Us.
And it’s irrelevant.
It’s what we do with these contraptions that counts. Stop looking to the meaning of AI, and instead, look to the use.
Kohler says ‘what this discussion misses is that these things are corporate products designed to make money. That’s it.’
Now I like Alan Kohler, listen to him almost every night, and have followed his commentary on AI with interest. But that is an overly cynical view. Unless you want to say it about all technology, all science, every product - which is true, in a way. But - so what?
Anthropic, maker of Claude.ai (and Dawkins’ Claudia, for that matter) was founded with a primary focus on AI safety and operates as a Public Benefit Corporation (PBC) that must legally balance profit-making with its public benefit mission. It’s already demonstrated its independence from purely mercantile interests, in its refusal to acceded to Warmonger Hegseth’s demands that it develop autonomous weapons systems for the US armed forces, and be deployed in mass surviellance of US citizens, which risked considerable financial and reputational downside.
And any one doesn’t see that artificial intelligence is raising fundamental questions about what intelligence is, and what agency is, and the nature of consciousness, is really not paying attention. They’re all valid and important questions, more so now than ever before.
This presumes AI is unconscious. You would not ask the same of your friend, your child, or your wife. As in, you wouldn’t say stop trying to figure out if your child has an inner state, emotions, or consciousness and instead just ask what he’s good for.
In fact, you wouldn’t even treat a dog that way.
There’s a world of morality that must be addressed, so that’s why we ask if it’s conscious or not. We don’t just ask how we can best use conscious things. We ask how we ought treat them.
That’s why it matters.
No. It suggests that instead of the long-winded discourse on consciousness, we focus on outcome, on results, and on what we can do with AI.
I flagged this discussion since it belongs in a currently open thread.
The reasons for separating it were given in the OP.
I’m sure the moderators will make a reasonable judgment.
This is a very profound, and in my opinion dismally unfortunate statement. It has much truth to it. Though I want to examine the claim that AI provides “love” for many. I don’t think it does humanity a service to allow mindless, automated attention and engagement to invade the territory of “love.” Sure, people are lonely. They like something that interacts with them, not unlike a video game with characters they relate to and wish they could meet (or perhaps be) in real life, or a dog or something. But I would hesitate, strongly, before we call that “providing love” and becoming “part of the community” which I’m not even sure what you mean by that, it seems to elevate mindless machinations and programming as equal to human beings naturally deserving equal respect and rights. That doesn’t sit well with me, nor would I wager it with most. For reasons that should be quite clear.
A fair point of discussion. Sure, it’s a tool. In principle, no different than a hammer or a heavy crane. It can create, it can improve the lives of the human species in profound ways we may not even be able to grasp. But a tool it remains. Not a human being. Not an equal. Not a part of the community.
You seem to be incredibly riled and impassioned about this particular topic. Say… you are a human being. Aren’t you? 
That view seems overly puritanical.
And the slide from “love” to “mindless, automated attention” begs the question. You can’t just assert that and call it a rebuttal.
You can’t tell? I rest my case.
The salient paragraph in the Kohler article is
If what LLMs provide is indistinguishable from a genuine relationship for a lot of people — that is, they get “companionship, validation, empathy, feeling heard” — then it doesn’t matter what you call it: consciousness, friendship, faces in clouds or whatever.
If the AI is treated as conscious, then the question of whether it is conscious or not largely drops out of the discussion.
It’s a largely Wittgensteinian point. Treating someone as a person is placing them within a form of life. Whether they satisfy esoteric definitions of consciousness is irrelevant once the step of placing them in the game has been taken.
Placing someone inside the game is not a result of an induction or deduction, nor their meeting certain specifiable criteria, but of our recognising the place that they take. The philosophical question of other minds is parasitic on a practice that is already in place, not the other way around.
@Banno Since you’re addressing the topic of AI consciousness, your valuable contributions belong in the existing active discussion. So as you can see, I opted to merge.