Our fourth episode discusses the concept of the “possible” and provides a brief argument for why we do not think AI is possible.
[Please note that this may not exactly match the audio. However, there should be no significant differences].
Welcome to episode four of the Metaphysics of Physics podcast. I am Ashna, your host and guide through the hallowed halls of the philosophy of science today. Thanks for tuning in!
With this show, we are fighting for a more rational world, mostly by looking through the lens of the philosophy of science. We raise awareness of issues within the philosophy of science and present alternative and rational approaches.
You may want to subscribe via iTunes or any of our other subscription methods. Also, you can follow us on Facebook and Twitter. You can do all of this from the shownotes or the media player on the website, at metaphysicsofphysics.com.
Today we are going to discuss the concept of “the possible” and some reasons why we do not think that it is reasonable to believe that AI, by which I mean computer AI, is possible.
Ok, let’s get into it.
Let’s establish the proper meaning of “possible”. I think a lot of people, including a great many in science, do not grasp what this word means. Something is not “possible” simply because one cannot think why it is not true.
“X is possible” means: “That according to the context of my knowledge, there is some evidence that X is true and none that proves that it is not”. Note, that there may not be very much evidence that supports X, but there has to be at least a little bit.
The evidence might not be very conclusive, and the truth of X may still need to be verified. But, at least I have established the possibility of the claim.
Let’s take an example.
Suppose that I come to understand that some organisms seem to undergo a process of evolution by natural selection.
Let us also suppose that I do not know how humans came to be. I wonder if humans might have evolved from some other species. I have not yet established that this is possible, at this point, it is just speculation.
So, I then proceed to find evidence that would suggest such a thing, such as perhaps fossils or genetic evidence. The evidence seems to suggest that man evolved from some earlier form of man-like ape.
Therefore it seems possible that man evolved from an ape! I have not simply asserted that it is so because I can imagine that it is so. I have concluded that it might be so, on the basis of some evidence that seems to suggest this.
Am I certain of this? No, all this is not very conclusive yet. But, I have some reason, based on evidence, to think that it might be true. I have a basis for my belief that man evolved from this earlier species.
But, suppose that I simply declare that man evolved from apes. I have no evidence of any kind to suggest that this is true. I just assert that it is so because I simply believe it to be so.
This is speculation. It is not the same as establishing that something is possible. You have no reason, based on no evidence, to believe it might be true. One cannot base an argument solely on speculation.
So, for something to be possible, you have to have a reason, based on evidence, to think that it is so. It also has to not contradict known facts, as we shall see in a moment.
According to this definition of “possible”, it is not true that “anything is possible”. Unless you can establish a reason to believe something is true, it is not possible.
Therefore, lots of things are not possible and cannot be established to be so. Note, that all gods fall into this category. No valid evidence of any god has ever been given. By their nature as allegedly supernatural entities that allegedly supersede nature and thus evidence, no evidence can ever be given for their existence.
If you can establish if X was true it would contradict a known fact, then X cannot be possible. There cannot be any contradictions in reality. Nothing that is true contradicts something else that is true. If you know Y is true, but X contradicts Y, then X is not possible.
Let’s say that I wish to establish that man is a reptile. But, I know that man is a mammal. But, if man was a reptile, this would contradict the known fact that man is a mammal. Therefore, it is not possible that man is a reptile.
I think that just about covers the issue of the possible. Let us move on.
If anyone wants to claim that computers can be conscious, or intelligent, then the burden of proof is on them. Nobody in the history of AI has ever passed the most basic hurdle required for such proof.
Having said that, it is easy to establish that computers cannot be expected to be conscious and therefore they cannot be intelligent.
They can only do what their instructions tell them to. Otherwise, nothing happens. This is a fundamental fact. This is a fundamental aspect of a computer’s nature. It is simply a bunch of logic gates, on and off switches and it only does what its instructions tell it to and nothing more. It cannot do anything else.
A computer, at least when you consider the parts that do the processing, is literally a bunch of on and off switches wired together. A computer toggles these switches on and off according to a pattern determined by the programming code. The code is first reduced to what amounts to a combination of “on” and “off” instructions, which are usually represented as ones and zeroes.
The programming code simply tells the computer which switches to turn on and off. That is essentially all it does. It seems amazing that simply toggling switches on and off is as useful as it is, and it is amazing. By toggling these switches on and off we can store all sorts of patterns of things and cause all sorts of things to show up on monitors, printers and so forth.
But, it should not be forgotten, that as magical as all this seems, when it comes to processing, you are essentially just toggling whatever switches the programming code tells the computer to.
Ah, some would say, computer programs do amazing things. They seem to be able to beat people at chess and predict what people want to buy on Amazon. And other things that seem to require an understanding of things and the ability to think and perhaps even reason. It sure seems to understand what is going on and to be able to think!
While humans require intelligence to perform certain tasks, computers do not. Just because a task requires intelligence for a human to perform it, does not prove that a computer must be intelligent to perform that same task. All that is required for the computer to perform that task, is that it is possible for the task to be performed by executing programming code. A computer can be instructed to do many things which do not require the computer to have any intelligence.