Episode Four – The Possible and AI

Play

Yes, a computer can be used to beat people at chess, help make predictions and many other interesting things.  But, what these things imply is intelligence on the part of the programmers.  The programmers are intelligent enough to know how to instruct the computer, step by step, to toggle the right switches so that as a result someone loses at chess or sees something on their screen that they might want to buy on Amazon.

All that is happening is that the computer is taking in input, in the form of on and off commands, and then processing that input.  The processing of that input is simply toggling switches according to the instructions given to it. It is simply generating output. It is turning pixels in your monitor on or off, it is printing dots here and there on paper, it is sending signals to other devices and so forth.

Readig printout
Here is someone reading meaning into some output.

Obviously, if you want to do anything useful with a computer you have to understand the output and to be able to do something with it.  You might look at the output of your “understand consumer behavior” program and interpret that as an indication that your consumers will buy this product but not that one.  But, that is reading meaning into the output.

It might be true that you can do amazing things with the output, but you must not confuse the intelligence required to construct the computer programs and use the output with any kind of intelligence on the part of the computer.  It has none, it is simply blindly doing what it is told, without any awareness of the problem, without any ability to consider the problem, or any awareness of anything else.

This is not artificial intelligence, it is human intelligence being employed, with the computer being used as an unthinking tool.

At best, computers are an aid to human, natural intelligence.

Now, let’s get onto consciousness.

Intelligence requires consciousness. If you are not conscious of anything, then you are not aware of anything and you cannot think or know anything. This would be like saying that a brain floating by itself in space or otherwise without any sensory input, could have anything to think about.

How is it going to think or know anything? Think or know about what? There is nothing about which it can think or know.

Now, tell me how you generate instructions for consciousness and if you tell me how we can consider intelligent computers remotely plausible. Give me any reason to think this is possible. Until then, you cannot even pass the most basic burden of proof and you have failed to establish that AI is even possible.

Note, possible does not mean “I cannot think of a reason it cannot be so”. Possible, means “I have a reason to think this might be true, based on the context of my knowledge”. That is, you must have a reason to think that it might be true. Belief without reason is speculation and you cannot base an argument upon speculation.

That is a faulty premise held by most advocates of AI. They cannot think why computers cannot be intelligent (even though it is fairly simple to understand). And so they think that “I can’t see why not” is the basis for thinking something is possible.

But, let’s suppose a computer had consciousness. You still have a problem. Unless you give it volition, it still lacks intelligence. Without volition, it still cannot do anything other than what it is told to do.

Why does a computer need volition to be intelligent? Intelligence is not simply the process of blindly executing only the instructions contained within computer code.

It requires the ability to grasp the facts of reality and to exercise one’s mind in order to deal with them.  This requires the ability to choose to focus on problems and to develop the ability to think so as to gain knowledge.

All of which is impossible without volition.  And, as we have established, without consciousness.

So, you also have to find a way to generate instructions for volition. Good luck with that.

It is not reasonable to believe that you can replicate consciousness or volition using any combination of programming code.  Certainly not without establishing the possibility of either.

AI
Sorry, there will be no HAL. That might be a good thing …

Well, what about emergence? Is it possible that intelligence is an emergent property of combining computer technology together in the proper way?

Saying “I think that AI is an emergent behavior” adds nothing to the discussion, really. Unless you can provide any evidence that intelligence could be an emergent behavior, then it is simply a blind, arbitrary assertion.

It is easy to say something is emergent.  But, that once again is assuming something is possible. This time because you believe that establishing something to be complicated means that you can get unexpected results.

Yes, it is possible that some systems can exhibit unexpected results.  But, all that proves is that you have not fully understood the system. If you understood the system well enough, then it would not be unexpected.  You would know that certain kinds of behaviors should not be expected and that some might even be impossible.

So, unless you can show that intelligence might be an emergent property of computer systems, it does no good to assert it might be or is.

Claiming emergence does not excuse you from the burden of proof.  You still have to show that consciousness and volition are possible to computers.

Alright, I am just about done for now. Before, I wrap up though, a few more things.

If you have not done so already, please check out the website on metaphysicsofphysics.com.

The show will officially be moving to a fortnightly schedule, at least until the start of the new year.  Content will be released every two weeks, although, in some two week periods, more than one episode might be released.

You are welcome to send in questions about any of the things talked about in this episode or about irrational stuff in physics or the philosophy of science in general. Send them into: questions@metaphysicsofphysics.com.

Join us again here in two weeks for the next episode in which we will be covering some frequently anticipated questions and some other issues.

Thanks for listening! Please tune in for the next episode and start thinking of some questions! Until then, stay rational!

2 thoughts on “Episode Four – The Possible and AI”

  1. One possible way in which AI may emerge is the continual replacement of human parts until there is no longer any organic parts left.
    e.g. as of today I can replace most parts of a human, legs, arms, heart, most organs etc..
    On the head I can replace the eyes, ears, nose & some parts of the brain.
    As we understand more of what it is to be human we will be able to replace more of the brain.
    Eventually, the “consciousness” part of the brain will be replaced & on that day we will have an artificial AI or artificial human.
    So like organic evolution I think artificial evolution will occur in steps over a significant time period, but significantly less than that required by organic evolution i.e. 1000s of years not millions
    Note: Given the expansiveness of the universe the only way humans can explore it is to evolve into artificial bodies since cosmic radiation is lethal to organic life & time travel & FTL travel are an impossibility

    Reply
    • Interesting thoughts there. I agree with some of this but not so much others.

      I am going to cover this on the next podcast, due next week. So tune into that if you want to hear my thoughts on this interesting and evidently thought-proving comment. 🙂

      Reply

Leave a comment