Thread: Could a computer program have a consciousness?

And then we reach the question of what distinguishes something shaped like a neuron from a CPU, in any way that has a relation to "consciousness"...

Originally Posted by RedfishBluefish

And then we reach the question of what distinguishes something shaped like a neuron from a CPU, in any way that has a relation to "consciousness"...

If they both produce consciousness, then it doesn't matter.

Originally Posted by Serkat

I meant to say "neural in nature", i.e. "nature" not referring to the actual chemical quality of the material, but the macroscopic organization of the system's parts. So, artificial materials yes, but organized in a way other than RAM+lots of conductors+CPU, running a neural simulation software.

Our brain is organized in a way that is equal to that artificial neural network. I don't see why it matters that a CPU is doing the processing of the data.

A better way to look at an artificial neural network program is to see the CPU of the computer as a heart that is just keeping the program alive. When you go all the way from CPU and RAM to the actual program, memory and CPU power has been abstracted so much that the hardware has absolutely nothing to do with the function of the Artificial Neural Network(unless the network is programmed in a low level language).

Originally Posted by Pensive Patrick

If they both produce consciousness, then it doesn't matter.

Precisely.

Originally Posted by RedfishBluefish

And then we reach the question of what distinguishes something shaped like a neuron from a CPU, in any way that has a relation to "consciousness"...

Well, a CPU is centralized and it does very basic computation. Neither a brain nor a neuron compute things or work on a digital foundation. I have no clue how CPUs work though, but I'm pretty certain that it can be explained by one of those diagrams that have 50 boxes in it and 2000 lines and arrows that are friggin confusing. But I guess it's pretty different from a biochemical neural network.

Originally Posted by ClassyElf

Our brain is organized in a way that is equal to that artificial neural network. I don't see why it matters that a CPU is doing the processing of the data.

Well, because "data" as such I see as objectively non-existent. Using "data" to model the brain might be fairly accurate, but it's still just a model. The brain doesn't work in terms of "data", unless you interpret it as such. The whole thing is more of a process as a whole than a process that uses an "object" (data).

A better way to look at an artificial neural network program is to see the CPU of the computer as a heart that is just keeping the program alive. When you go all the way from CPU and RAM to the actual program, memory and CPU power has been abstracted so much that the hardware has absolutely nothing to do with the function of the Artificial Neural Network(unless the network is programmed in a low level language).

Yes, I know this. But again, I don't see a "program" as something that exists anywhere else but in social construction. I don't think that hardware can ever have nothing to with consciousness. Sure, software might use neural networks but that doesn't change anything at all about the actual state of things which is the same ole extremely basic digital computations in the CPU that are used for any other type of software.
I.e. software can't be conscious, only matter can.

If you argue that a "program" can be alive and/or conscious, you might as well argue that a book without a human is actually more than ink on paper... that the information is floating around in some universal off. Or that Super Mario actually exists somewhere some time if I pop the sucker into my SNES.

Originally Posted by Serkat

If you argue that a "program" can be alive and/or conscious, you might as well argue that a book without a human is actually more than ink on paper... that the information is floating around in some universal off. Or that Super Mario actually exists somewhere some time if I pop the sucker into my SNES.

but Mario is just programmed to make pixels flash off on a screen under certain circumstances(button is pressed) . Neural nets go through a self examination of knowledge saved as grids of neuron arrays in order to decide what action to take.

Originally Posted by Venomblood

Humans are simply made up of atoms, just as computer programs are made up as bits and bytes. If someone were to (in the future) make a computer program simulating atoms and physics perfectly, then made a human out of those atoms, would that person have a consciousness? Both are made up atoms that can be seen and interacted with by their respective owners. Both can think, feel, taste, smell, see, hear. What defines us humans as different from a computer program person?

I had this exact same idea recently. And it is pretty cool: It is hard to formulate a reason why this simulation would not have a concious. If the simulation of atoms (and subatomics ect) is indeed perfect, and the human is a perfect replica as well (you would probably need to let evolution repeat within the simulation, or maybe you could just create in the virtual world a setting capable of letting a human get born by using someone's DNA-code). Grated this (hypothetical) perfection, there is no real difference, thus there will probably be as much concious in the simulation as there is here.

The fragmented time-passing of the simulation (actually fragmented creation of the simulation, since a computer is not capable of calculating everything at the same exact time) versus the (perhaps seemingly) fluent time-passing on our reality maybe be called a difference, but it would be hard to point out how exactly this changes things. From inside the simulation there is no real difference.

You could undoubtfully simulate a couple of neurons to let an robot-like creature get along in a virtual world, so you could -given enough processing power- simulate billions of neurons that all also work the same as in reality.

All this together makes it a greatly interesting though-experiment. It probably isn't going to overturn the bio-chauvinists, nor the stubborn dualists (people that believe thoughts/mind isn't mere substance, or doesn't originate from mere substance, but something non-material as well (note: something either completely undefined or completely unscientific)). However, it does shed an interesting light on the whole conciousness-discussion, if you ask me.

Originally Posted by ninja9578

You can't map atoms. Heisenburg uncertainty principal. You can either know where it is, or what it's doing but not both, and measuring it changes it.

No, the Uncertainty Principle puts a lower limit on the product of position and momentum certainty. So that limit (which I think is hbar/2) determines when the Uncertainty Principle becomes relevant. And it turns out, it doesn't really apply until you get down to individual electrons. Whole atoms can be pinned down quite accurately with respect to their size.

I found a research paper today on an AI named HAL.

It's the most interesting paper on AI I've ever read.

http://www.a-i.com/show_tree.asp?id=19&level=2&root=12


I am convinced that computers can become conscious after reading through this. Read the "First Words" page, it's amazing.

Computer consciousness: I think that's completely possible. Will we see that within the next 100 years, maybe not.