Turing Test

[PiMaster]

Recently my father and I got into a debate about artificial intelligence. He didn't believe we'd ever be able to fully simulate human intelligence, and even if we ever did, it wouldn't actually be intelligent. Obviously, I thought the exact opposite.

An interesting thought concerning this is from the classic book, "Xenocide" by Orson Scott Card (if you haven't read it, I recommend it). At some point in the book, Miro and Ender discuss free will, and if it even exists at all. Ender pointed out that if something with "free will" is created by someone/something else, the creator is almost like a puppet master. The master programs his puppets to behave in set ways. Computer programs are created by people, so they don't have any free will. He also pointed out that people are created by their environment, and therefore don't really have free will - it's only a script they play out.

This then begs the question - if we're ever able to simulate human intelligence, what's the difference between a "simulation" and the real thing? How do we know that the simulation isn't genuine, that the program isn't acting out a preset script given to it by the programmer?

[Bazzy]

Artificial Intelligence is strictly connected to randomness. The more randomness you can get, the better the AI will get.

This is the same for human brains, some thoughts can be the result of random interactions within the brain, and this makes the free will since it's not determined by the environment ( the interpretation of that thought may be )

Furthermore if the creator allows learning and intelligence evolution, the created intelligence won't be constrained to the set of situation that the creator may expect

[chrisname]

I've long wanted to create a program that opens it's own executable file, randomly selects a single bit, and flips the bit's value, and see how the program ends up. Obviously I would only let it select bits from the code and data sections of the executable.

[helios]

a program that opens it's own executable file, randomly selects a single bit, and flips the bit's value

I believe that's called a "carcinogen". You'd just be giving the executable cancer.

some thoughts can be the result of random interactions within the brain, and this makes the free will since it's not determined by the environment

But the person isn't in control of those interactions, so how is that free will? And the brain is not an isolated system, anyway; it's also part of the environment. Those random interactions are at least partly influenced by external conditions.
Randomness isn't the same as free will. If you take an intelligent entity and add randomness, you haven't made it free, just unpredictable. Free will, assuming it exists, implies an unpredictability beyond pure chance. A random event is uncontrollable, but the "will" part of "free will" means that there is something in control -- the one with the will. It's unpredictable not because it's random, but because its behavior is not dictated by outside influences; hence, it's "free".

if we're ever able to simulate human intelligence, what's the difference between a "simulation" and the real thing?

That question could be applied to any emulation. The answer is that perfect emulation makes the distinction entirely artificial. A program can't know for certain whether it's running on an emulator or actual hardware. To an outside observer, the difference is obvious. The behavior may be the same, but the origin of the behavior isn't. But is that distinction useful or even relevant?

I don't think we'll ever be able to create a strong AI. Not because it's fundamentally impossible (if one hardware-software platform can do it, why can't another?), but because it's a) hard, and b) impractical. Can you imagine what a "universal declaration of synthetic life-form rights" would look like?

[jsmith]

a program that opens it's own executable file, randomly selects a single bit, and flips the bit's value

That is, in a way, the theory behind genetic algorithms.

[chrisname]

That was what I thought. It was sort of like genetic mutation.

I'd only need a few hundred billion instances of the program (each with a separate executable file) and a few million years and eventually one of them would turn into something useful.

[R0mai]

@chrisname: Look into this, I think you might like it : http://stauffercom.com/evolve4/

[chrisname]

That looks fun. Downloading source now.

There was something like this I saw before but I could never get it to install on Linux... now I've forgotten what it was called.

[clover leaf]

I guess it's possible that you can get a whole network of computer, each one simulating a lobe of the brain.

p.s. If you like simulators like evolve4 I think you should google search "powder toy".

[chrisname]

How is the Powder Toy anything like Evolve 4? The Powder Toy is a Physics sandbox. It's also quite unrealistic (uranium isn't fissile) and has quite buggy source code. That's not to say it isn't a fun toy, or that you can't make cool things with it.

[helios]

Maybe it's assumed to be U-235.

[chrisname]

But all it does is generate heat under pressure. It's still pretty cool, and some other aspects of it are reasonably realistic. You can make nuclear explosions with it anyway, so it's good enough for me.