h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Alan Turing's test is all about belief. Some people already believe in AI, because they read the blurb that came with their latest kitchen appliance. Many people will remain unconvinced even after they have bought their first insurance policy off an AI, without realising it wasn't a human.

What of the AI itself, will it believe in us? Will it believe that we created it? Perhaps communication will be so difficult (with an alien intelligence) that we won't be able to make ourselves understood to the AI we create.

So, is there a God who created man? or does man create God.

Has anyone else read "The Jesus Incident" by Frank Herbert? This has a nice take on the AI/God thing, with a rather different perspective to the questions I just asked.

You have just suffered my first contribution to h2g2.

Thanx for the input, Deryk. These are all very valid questions/remarks. There's still a lot for us to consider when talking about the "God(s)" we pretend to know or believe in.

And what we are doing to ourselves and to any lifeforms we will be "creating" eventually? Well, maybe we should start feeling ashamed in advance. If I were an Artificially Intelligent lifeform, I think I would be as much of a non-believer as QT-1 (Cutie, by Isaac Asimov) was, regarding the possibility that maybe man created me.

There's still so much for us to learn...

Alan Turing's test was also conceived in the mid-fifties, when the hope of a lap-top would have been regarded as a Nostradamus prediction, rather than a viable possibility. Consequently, he couldn't account for the wonders of JavaScript, CGI or even Apache servers.

As long as computers still function from logical, ordered code they'll run afoul of humans who tend to react emotionally and sometimes think about it later. Part of the problem that many have with current computers is that people come from a place of wanting and expecting, and computers simply follow clearly laid-out instructions....regardless of how clever they appear to be.

Whether AI believes in us, to me, is irrelevant. The question is How much power are we going to cede to these very very fast and powerful machines? The more we entrust our lives to automated systems, especially of the AI variety, the duller our minds become and the lazier we get....serve us right if the machines rebelled.

If you're entertaining sf references, check out "Colossus" and "Colossus and the Crab" (AKA The Forbin Project); the reason why HAL went mad in 2001:; "I Have No Mouth and I Must Scream" by Harlan Ellison;.

Fruitbat

Sorry to try and hijack your debate with a touch of mysticism, it was just the mood I was in yesterday.

I do wonder what people think the point of an AI is though. Do we want something that can act as an intelligent agent, but without actually being a "living" entity that we had to care about? I think that the purpose of an AI is usually expected to be something along the lines of an intelligent machine to build cars/fight wars/talk to, but this presupposes that an AI would want to do these things for us.

There is a kind of dichotomy here between a machine that acts as though it is intelligent, ie. makes human-like decisions, and one that is genuinely an AI. Perhaps the difference is consciousness, but that is a long debate.

I happen to believe that consciousness (as we suffer it) is a consequence of a sufficiently complex system, and is certainly an attribute that exists on a sliding scale, not simply on or off. If there is a limit to what a human can understand, then I think that we will cease to understand our own creations at about the time they develop consciousness.

Alan Turing was pretty clever, I expect he put some thought into what could be developed if computer power/availability was not so restricted. Maybe he invented JavaScript, but couldn't see the point in writing it down at the time. He would be able to tell the difference between Deepblue and AI.

Ceding power is an issue, but I think it will come about through us ceding rights to machines. Today's intelligent machines have the power of decision that we have given them, but that power is restricted because we don't trust machines. By which we mean that we don't think the engineer thought through all the possibilities during the design. We don't, yet, have to worry about the machine's motivation. A genuine AI could have rights and responsibilities, like the rest of us.

Thanks for the references. I've seen Colossus, but I'm not sure if I read it yet. I read a review of the Harlan Ellison book. Unfortunately I don't have the time to read that I used to have.

> I happen to believe that consciousness (as we suffer it) is a
> consequence of a sufficiently complex system, and is certainly an
> attribute that exists on a sliding scale, not simply on or off.

We need more people like you around here.

I'm not sure about a sliding scale of consciousness, though. Emergent behaviour in complex systems usually exists in distinct "regions" seperated by boundaries or phase transitions. If, as I also believe, mind/consciousness is an emergent property of a suitably complex neural system, then, as nervous/brain complexity increases, you won't see a mind appear until the system reaches a "critical" level of complexity.

Having, written all that, I worry about what you meant by "sliding" and that I might have misunderstood. There may well be more than one of these phase transitions or regions of behaviour. That would imply a scale of sorts, though it would have discrete steps.

Well, perhaps the level of intelligence that humans have achieved is what we call consciousness, perhaps there are levels above this that we can't see (yet?).

As far as the question of consciousness, I meant a smooth sliding scale when I wrote originally. It seems to me that the mammals have a continuum of consciousness (ignoring the problem of the mice for a moment), from some of the hunters (cats, dogs), up through the simians to us, maybe with some aquatic mammals in the list somewhere. I don't see any real steps, although it is possible to measure certain functions, like the ability to recognise ones-self in a mirror for example.

Certainly in some complex systems there are steps in emergent behaviour, where the totality suddenly stops looking like a collection of all the little bits.

Wouldn't the "sliding scale" refer to the fact that we ourselves already show several different levels of consciousness? Such as "our self-awareness is different when we sleep from when we are awake", for example?

Another example would be: when we're pissed (loaded, blended, have little blood in our alcohol), we may still be aware of our existence, but we're not completely aware of exactly HOW and WHERE we exist.

So, you're talking about the "sliding to the floor" scale?

I wasn't even aware this scale had its own name

I'll have to think about this.

Here's something you might be able to help me with. When estimating the "intelligence" of animals, alive or extinct, biologists and paleontologists cite the ratio between brain mass and body size - "X had a very large brain for it's body and so..."

If we assume that a neuron is a neuron and is pretty much the same size throughout the animal kingdom and that mind is an emergent property of brain activity, why consider body size? Won't absolute brain size do?

It's proper name is the Catatonic Scale

Paleontologists cite the ratio between brain mass and body size because it is all they have to go on, copies of The Times crossword don't fossilise too well. The justification is that the brain needs to be a certain size to control a large body, although when I think about it now this may be just another fiddle factor they have made up. I think there is some sort of theory about some large animals with small brains eg. a stegosaurus, requiring a "second brain", more like an amplifier, half way down its body to do the control.

For consciousness or intelligence I think absolute size has a lot less to do with it than the complexity. Complexity just maps to size in brains because mammals all use the same technology in their skulls.

No, I guess it won't.
The problem is: the larger the body, the more brains is needed to properly motorize it. The brain need more complexity just to precalcultate the right amount of propulsion, and possible actions in case of emergencies.

However, I *would* agree that a normal ratio-function is not fair.
For example: a creature twice our size would not necessarily need twice as much brains as we have, just to move around. Maybe 25% (or less) extra would suffice.

That sounds like the basis of a good guide entry to me.

What I was actually referring to, is this:

Our brain has reached such a level of complexity, that it needs several levels of awareness, just to handle all the input. Each level handles the input in a different way. Some things will even have to wait untill we sleep, before it is processed.

> The problem is: the larger the body, the more brains is needed to
> properly motorize it.

Why? I wouldn't expect to see this. A seven ton T. Rex would have pretty much the same muscle groups as a modern domestic cat. Why wouldn't a brain the size of a tabby's be just as intelligent in a animal the size of a T Rex? The latter's muscles would be bigger but they would require no more motor neurone input than the former's.

Maybe.

We need a biologist.

This sounds about reasonable. Maybe we could say that a certain absolute quantity of brain-mass is needed to get enough complexity for awareness to arise, but above that quantity something is needed -proportional to the mass of the body- to control motion. (*perhaps*)

This complicates matters slightly, because if we were to agree to the above, how would we weigh both factors?

What is the size of the "critical mass" needed for awareness and/or intelligence to arise?
and subsequently: How big (or small) a part of our own brain-mass is really necessary for us to control all motion?
What function do we apply to that mass, to calcultate the brain-size any other mammal would need?

As long as these questions ar unanswered, what else do paleontologists have to go on, except for the actual size of the brain?

Hmm, if that were true, then no mammal would need more brain-cells than any mouse has. Nature can't have been THAT wrong, can it?

Ah, so *that* is what this is all about? *grins mischievously*