Q&A: Consciousness in Artificial Intelligence
Consciousness in Artificial Intelligence
Question
Can artificial intelligence have consciousness? And is there any kind of test that could check that?
Answer
My inclination is no. I do not know of any way to test it. The arguments and tests I’ve seen so far have not convinced me.
Discussion on Answer
Correct. But not only that; even a biological creature has no consciousness unless it has a soul within it.
I have no idea. At the moment I don’t see any way to test it.
Is there a connection between consciousness and free will?
Not necessarily. Consciousness without choice is certainly possible. Choice without consciousness is more problematic, because choice is a product of judgment.
If I may throw in my two cents, I’d add this. First of all, you asked about consciousness, whereas the more precise question about the nature of the difference between conscious beings and a machine should also take into account the principle of self-consciousness.
Further, the dualist assumption (the existence of a soul alongside a body in a human being) is based on a complex consideration and, in my view, a justified one. I don’t see how one could use that same consideration to justify belief in consciousness (and self-consciousness) in a machine. It could be that in the future machines will be so highly developed that it will then be justified to wonder whether they pass the test, and we could include them too in our dualism. In my opinion, we are far from that.
Why is consciousness without choice certainly possible? Are there examples of such creatures that you know of?
Who said anything about choice?
I was responding to what Rabbi Michael wrote. As for what you wrote, Doron, what future behavior could you expect from machines that would make you think they have a soul?
The short answer, or at least one answer, is that if machines develop a material and spiritual history like human beings, we might perhaps think they have something in common with us (or actually, and maybe more plausibly, we might think that we are basically like them—sophisticated machines, and that we have no real “soul”). For that, those machines would need to make a bit more effort than they do today: build religious institutions, courts, make art and economics, fight with one another (or with us), produce offspring, create science and literature, and in general do lots and lots of nonsense that doesn’t fit with the algorithms driving them. When all that happens, I might become convinced that materialism is correct, and in that case the concept of consciousness would have no meaning.
Ahh… and maybe also if they succeed in creating human beings in the lab who are identical to us. The kind who don’t really believe that machines have consciousness.
How could I know such creatures if there’s no way to know who has consciousness? It follows from conceptual analysis. There is no necessity at all that whoever has consciousness must also have choice.
I’ve heard quite a bit from you about what free choice is. But how do you define consciousness?
You asked me questions about consciousness. So you define what it is. What were you asking about?
Why assume that artificial intelligence would have consciousness? Who said consciousness is connected to complexity? Can anyone put a finger on the stage that necessitates the creation of consciousness? And if not, then why tie them together?
Papagio,
In my opinion, the way you presented your question is incomplete.
Broadly speaking, there is no philosophical or scientific question about the very existence of consciousness itself (everyone agrees that it exists or can exist). The main debate is between materialists and dualists, with the former arguing that since everything is matter (and energy), human consciousness must be explained only in those categories. Accordingly, human consciousness will be defined by them only through its functioning and its products, while denying the existence of a metaphysical and “spiritual” substrate that supposedly serves as a kind of anchor for those functions and products (in the eyes of their dualist opponents).
As for your question—where exactly to put one’s finger on the stage at which consciousness comes into being?—it seems to me that the answer splits according to the two camps I mentioned: the materialist will assume that consciousness is inherently intertwined with matter; of course, what is meant is a very complex and highly sophisticated material object like a computer (though it may be that we are talking only about the computers of the future…). The computer’s “consciousness” came into being—or will come into being—only when it was created. At the very least, it came into being as a potential (think, for example, of a computer that has just been manufactured in a factory but has not yet been connected to electricity—it has only potential consciousness).
By contrast, the dualist will have to assume, in my opinion, that consciousness (or at least the substrate that bears it) can exist even without matter. That is definitely a clue regarding its supra-temporal status (though not necessarily matter itself). According to that approach, it is not bound up with matter.
Hi Doron.
1-If so, it turns out that according to the dualist there is no room for discussion at all, because there is no logical connection whatsoever between the assumption that there is consciousness and the functions that matter performs even without it. And therefore there is no reason to assume that artificial intelligence would have consciousness. (By the way, there are things in which computers are far more sophisticated than a human being even without consciousness.)
2-By the way, for a long time now I’ve had what seems to me to be a crushing argument against those who claim that consciousness arises from matter: insofar as you assume that consciousness exists but cannot explain what function it serves (since one can program an intelligence to perform any function a human being is capable of), it follows that consciousness is not connected to matter and is an abstract spiritual entity. What do you think?
No room for discussion? That’s a bit much. I’m a dualist, but I could always be wrong. Maybe far-reaching technological development or philosophical considerations I haven’t thought of will change my mind. Why shut down the discussion in advance?
It seems to me that here you’re talking about what is called emergence, a position that also seems flawed to me. The various emergentists assume that one substance, spiritual, can somehow emerge out of another substance—material. But then they have to have it both ways: somehow properties that belong only to spirit (like intentionality) somehow exist potentially in matter, and in some mysterious way they manage to be actualized only outside matter (that is, in consciousness). I admit I haven’t gone deeply into the details of the emergentist position, but it seems to me that this is its main thrust.
If consciousness is what we call knowledge (or awareness—more precisely, a collection of items of knowledge that form a complete web, creating awareness of the conscious subject as an entity separate from the rest of the world), then it also requires free choice. For there is no knowledge or awareness without prior contemplation. And contemplation is precisely a matter of choice. A person has the choice whether to contemplate or to “shut” his eyes. Mere sight is not consciousness. Animals have that too. For what the eyes see to say something to someone, there has to be some learning process that turns sensory data into concepts and items of knowledge; that is contemplation, which, as stated, is a matter of choice.
Rabbi Michi,
I assumed you had a definition of consciousness, because you said that consciousness without choice is certainly possible on conceptual analysis, which implies that you can define both concepts.
In any case, I think that for a creature to have consciousness, there has to be some kind of “I” that can relate to and recognize itself. I think that this “I” must be external and not completely controlled by the deterministic system, because it sounds contradictory to me for a system to recognize itself. Assuming that is true, there may be a connection between consciousness and choice, if this external “I” (which perhaps one could call a “soul”) is what controls the physical creature.
In any case, it’s a slippery concept, and it doesn’t seem to me that I’ve invested enough effort to define it properly.
Why are you inclined to think not? Because it isn’t a biological creature? Or because it isn’t a human being (and has no soul)?
And do you think there is theoretically some test that could be done, even if you don’t know it? Or would we say that no matter what happens and what is discovered, we would conclude that it has no consciousness?