Is it possible to make Intelligent Machines?

Intelligent Machines

I have had some discussions with engineers who had Intelligent Machines as their religion, and interestingly it’s the Atheists who dreams about thinking machines.

With background in the current computer technology I will analyze the feasibility of intelligent machines from three different scientific approaches: The Mathematical, the Physical, and the Biological.

It’s interesting that the believers in Intelligent Machines ignores the evidence, especially the stringent Mathematical evidence, this is also the reason I call them believers as they ignore facts, dreaming of a future with god-like machines a lot wiser and knowledgeable than man.

Mathematical Evidence

Kurt Gödel made the Incompleteness Theorems:

It is possible to have a complete and consistent list of axioms that cannot be produced by a computer program (that is, the list is not Computably Enumerable).

And

The incompleteness theorems also implies that not all mathematical questions are computable.

It states simply that there are problems which man can solve and machines can’t.

What he proves is that there are things in a system, like our physical world, that can’t be done within the system itself, which can only be done from something outside that system, but man breaks this barrier where computers can’t, and that means that man depend on something outside the physical world, like soul and spirit.

The BBC program `Dangerous Knowledge (part 9/9)’ looks at this Mathematical Prof. As the video is destroyed at Youtube from time to time there is another source here: Human and computer knowledge.

Other documentation

Roger Penrose shows with examples why they cant think in: “The Large, the Small and the Human Mind“.

In this Hard Talk interview about the recurring big bang Roger Penrose touches the problem.

Physical Limitations

When you look out into the room, you are sitting in, you have a picture or really a video of everything at once and constantly. It means that this view of your surroundings, and your thoughts about it, are in your Consciousness. Seen from a computational view it is really fantastic, and Computer Science today can only dream of something that seems to be Conscious! Even with future technologies as Quantum Computing is it a question if we can make Conscious Computers at all.

Computers are something called Von Neumann Machines, which again is defined through the theoretical Turing Machine, and Turing Machines can only work on one bit at a time, which is only a small part of the information necessary to make a single point on a computer or television screen.

The consequence is that Computers can only have a part of a point of a picture in what we can call their ‘consciousness’ at a time, and that is not enough to even contain the color information of a point. What they have calculated in one instance is forgotten in the next. It’s only when we see the result on the screen or on paper that it becomes conscious through our consciousness, that is, we see the  screen as a whole, extracts the relevant information in all its complexity, and understands it’s implications.

A computer cannot be Intelligent however big and speedy it is.

Brain Science

Erich Harth in ‘Windows on the Mind’,  1982:

The brain presents two seemingly irreconcilable aspects: It is a material body, exhibiting all the physical properties of matter, and it possesses a set of faculties and attributes, collectively called mind, that are not found in any other physical system.

In his book “The Creative Loop,..” he elaborates further on how the physical mind functions, and why it’s superior to any known devices.

Literature

The book by Erich Harth’s “The Creative Loop, How the Brain Makes a Mind” gives an intelligent description of the physical working of our brain.

The book Stairways to the Mind by Alwyn Scott gives a broader overview of the area.

No Theory of a Conscious Machine yet

There have been talked about thinking machines for forty years, but there have not been a single theory for building one yet. There are a lot of programs running on super computers which can simulate some aspects of the mind, but the programs have nothing to do with real intelligence.

Computers are good at those things where we are bad, that is remembering and calculations, but they can’t think.

With the knowledge we have to-day, if it had been possible to program consciousness into our current computers, we would have done it by now.

Future possibilities

It may be possible to make thinking computers if its build on Quantum Computing, as the numbers of bits in the computer’s processing unit can be increased considerable above the one bit we have today, caused by entanglement, but I don’t think that it’s enough to create consciousness, but it is a requirement.

Literature

Hubert Dreyfus: What Computers Can’t Do: The Limits of Artificial Intelligence.

Quantum Consciousness: A Discussion between Stuart Hameroff and Alwyn Scott.

Tarjei Straume on the Technological Singularity.