RE: The Turing Test and Machine Intelligence (part II)
You are viewing a single comment's thread:
Yeah the key problem is that we can't have direct access to the consciousness of others. Nevermind robots: scientists are debating whether certain animals have consciousnesses. Do cockroaches have a consciousness? The question enters my mind every time I spray one and see it writhing. Many religious people don't even believe that other mammals, like dogs, have a consciousness!
Since consciousness is an inherently private thing, I doubt it can be detected using the traditional ways we use in science to investigate matter.
The closest I have come to figuring out a real test of consciousness is the following. Suppose scientists slowly shut down parts of my brain. How would I, subjectively, feel? Would I feel that I'm becoming less conscious? Alzheimer's sufferers show that probably I'd have no clue. But what if they then started turning on those parts again? I think I would know that, compared to before, I'm more conscious now. So, similarly, perhaps, let's say we have a robot brain. We could shut down parts of my human brain, and turn on robot parts that we connected to my brain, and then reverse the procedure, and compare how a person felt when he was becoming more and less robotic.
In principle, I see no reason why a future science couldn't get us back and forth through some mental state. You can probably give me drugs and make me super angry, or super calm. Since everything is matter, why can't we slowly make a human brain into an animal one, or a robot one, and then back again? The person would probably know or remember whether he was becoming less or more conscious during the procedure.
You get the point. It's sci fi stuff, but I can't see why, in principle, it couldn't work and, more importantly, I can't see what else could work! I definitely don't think Turing's test works. It wouldn't be much different from male beetles thinking bottles are female beetles.
Yeah, your take on consciousness is similar to solipsism. On the other hand, while it is true that some organisms, such as cockroaches and amoebas, can have a level of awareness of their environment that can even trigger an apparent "panic attack", I think that is still different from being conscious. I am more inclined to think like David Wallace: consciousness seems mysterious because it is complicated and we have not yet understood it fully, but it is still a physical process and we should not need special theories to explain it.
Yes, what we don't know is whether that conscious machine can be based on discrete states and algorithms. Turing had already imagined that creating a conscious artificial being of some kind must be possible since we exist, we are the living example. The problem is what should be the minimal nature of that machine. Our brain is analogical and we don't know yet if thinking is computational. Maybe that is the limit or maybe not.
Thanks for the discussion!