mway wrote:
Just a different brain state.

mway wrote:
As for the Chinese Room experiment, it is simply naive.

mway wrote:
The human in the room "follows" the computers code. Now replace the computer code with a neural map of a brain that can [obviously] pass the turing test. The human (who does not know any chinese) can perform the exact same task as that done with the computer code, producing the correct chinese characters. In other words, you can process a human neural topology in the exact same manner as you would manually process a computer program.

Kamerynn wrote:


So, suddenly, the fact that one can act as if does not presuppose dualism. That wasn't so hard, was it?

None the less, you haven't really answered the question. How do you know that brain state X is the pain state and state Y is the acting as if state? What's the difference?



This shows that you don't quite understand the implications, yet. It shows how a system can act as if it understands Chinese without actually understanding Chinese at all. That is, syntax doesn't amount to semantics. All a computer has is syntax. For example, that it can actually be in pain, and not just act like it, is something that you are assuming.



Exactly - the Chinese room is analogous to a computer program acting as if; they can both perform the same task, and both would pass a turing test. That the Chinese room passes the test without any implicit understanding of Chinese shows that the turing test isn't sufficient (only necessary) for actual understanding.

All of this is academic, though, if you can't see that the difference between a person acting as if she's in pain and her actually being in pain is the pain. That we feel pain isn't up for discussion. What is substantive, here, is how we can know that a computer is in pain and not just acting as if. Indeed, how can we even know that other people aren't just acting?

mway wrote:

Could you elaborate on that?

ssu wrote:
What I always find in these debates problematic is that we don't actually even know or agree on what consciousness is. What and who is conscious and who isn't? Us writing here hopefully are conscious. But is a dog conscious? If on "some level" that would be the case, then were do you put the line? A creature made of a couple of cells cannot obviously be conscious.

One typical way to look at this totally disregards some fundamental questions. The person who started this thread said: "Unless AI machines have exact copies of the wetware in the human brain, I doubt they will ever be conscious and obviously, we will never know for sure." But why would it be so? Just the same hardware in place and ZZZAPP!!! You have consciousness? Really??? In my view usually far too much emphasis is put on the structure how our brain is constructed. We tend to look at cells, genes, molecules, atoms or the latest physics-fad or math to give an answer. It is as if in some part of our brain is a spot from where we get consciousness. In short, the filosophical view of materialism, that matter is the only substance, limits the way the problem is looked upon. Why? Because with the usual materialistic world-view the focus is at matter and the importance of interaction of matter is neglected as interaction itself is not matter. But consciousness is all about interaction. (Note that this problem doesn't even relate to the question of is materialism true or not.) In my view we become and learn to be conscious, we aren't from the start conscious.

We can even see this when we look at how we use in other uses the verb "to be conscious". One becomes conscious about something (like global warming) by learning about it, before that one is ignorant. You read a book, watch a TV documentary or hear it from someone. And this learning is done by the use of language, in spoken or written form. And language is not something made of atoms and there lies the problem. It really isn't the ink on the paper or the movement of air when we utter a word. Sounds obvious, but before thinking that this is quite irrelevant, think really about it the technical problem with AI where an engineer has to put the "nuts and bolts" into action. The use of language and learning. Extremely important things to our ability to be conscious (or more conscious lets say animals). Now the engineer has to come up from matter, nuts and bolts, something that uses something that is not made from matter. The "Chinese room" argument clearly shows that language is not merely a simple interaction of different signs.

My personal view is that the problem of the consciousness AI is something comparable to the following: A computer is "conscious" if it can re-write the software that controls the actions of the computer itself in a way the programmer (of the software) did not program it to do in the first place. Hence it can make its own conclusions, not the conclusions of its programmer. Of course the programmer can make a computer rewrite it's software from outside stimuli (and this is said to be learning), but the software has to have a guide how this is done. The focus is thus on the software part "how to rewrite this software". For somebody with independent thoughts this is easy: you just look at what the program says/does and do something different. But for the engineer writing the software the "do something else than in this software" is a bit more complicated.

And that is just one problem...