Ai

[c_barnett01]

What do you think ai's are be like
e.g.
the computer of gene rodenbery's andromeda
or somthing else

[Olio]

well, i think AI are going to be like the robots in asimovs foundation series. but i may have spent too many hours up lately :)

[BountyX]

AI's will always be artificial as long as were the ones making it. In the far furture AI will be the closest thing to the human brain (by then i bet computers will be molecular instead of electrical since molecules accept sequences in ABCD opposed to just a 0 or 1;thus, faster), ofcourse we have to fully understand our brain first.

[doubleglazing2]

Hi,
Hmmm...interesting topic:cool: . I think most significant thing we can say about AI is that it will never be like us. Sometimes I start forgetting that we're not just grey matter - recently watched a documentery about a lady here in England who was air-head.

She had a massive hole in her brain -:idea: - so big that she only had thin layer of grey matter attached to the inside of her skull! But guess what? It doesn't affect her behaviour or intelligence in any way:!: .

I reckon our intelligence is itself half-artificial. We might even be able to emulate it in robots etc.. We're more than intelligence:-| , or emotion:cry: .

Thats what I think anyway....:) ....I love these smilies.. :cheesy: ..can't get enough pf them...lol

[Toba]

AI would be basically like a human being, but it would think differently.

I wouldn't really want an AI computer - but maybe I've been reading 2001: A Space Odyssey too much.

[Slade]

Define AI. If I were to say, emotion set us apart from AI, that would be bs since someone could just program AI to have emotion. The only thing I can really think of that would sprout a real argument is that we are spiritual beings and the mind and body are totally different entities. Anyway, just something to think about. :D

Slade

[Toba]

From what I've heard, programs are considered AI when they can fool a human being into thinking they are human (ie. over IRC chat or similar). Several programs have been able to do this sucessfully.

[K-1]

AI seems to be a limited scope of what it can be used for. Just the base AI, can only learn information, and link common data, make connections between words, things like that.

But how it's getting used is what the real question is.

When will dell switch to AI automated tech support.

[gusano79]

From what I've heard, programs are considered AI when they can fool a human being into thinking they are human (ie. over IRC chat or similar). ...

It's called the Turing test, after Alan Turing.

Here's an article that may be of interest... I'll throw in a quote from the article (italics mine): Turing intended his test specifically as a measure of human intelligence. An entity passing the Turing test may be said to be intelligent, and moreover, to possess a human-form of intelligence. Turing specifically states that the converse statement does not hold, that failure to pass the test does not indicate a lack of intelligence.

Much depends on how you decide to define "intelligence" and how exactly you go about measuring it.

Further reading:
http://tekhnema.free.fr/3Lasseguearticle.htm
http://www.pcmag.com/article2/0,4149,1200185,00.asp
http://www.google.com/search?hl=en&ie=UTF-8&q=Turing+test :)

... Several programs have been able to do this sucessfully.

Such as? Who wrote them? When? That's a pretty bold statement, and these "several programs" would have received quite a lot of attention, which I haven't noticed at all. And don't say ELIZA.

I'm of the opinion that if the person asking the questions is clever enough, he/she will be able to trip any program written to date into revealing itself as not a human being. I won't say that it'll never happen, but we're far from creating anything that will really pass that particular test.

--sg

[Toba]

Actually, I heard that from someone else. I don't know it for a fact, I suppose I shouldn't have posted it :( .

[Chainsaw]

One area of AI is pattern recognition. If a machine could reasonably understand AVERAGE people speaking commands, we could find that many (most?) service jobs would dissapear. All those McDonnald's cashiers could be replaced with a machine. Likewise, really reliable pattern recognition could make grocery checkout automated; put your stuff on the conveyor belt and it gets scanned and summed automagically. Construction machinery could be automated, and so on.

In other words, people wouldn't be needed in jobs that depend on our understanding of commands and visual recognition of items.

While I can see the up side of this, I personally hope it doesn't happen for a LONG time. Like the industrial revolution, many folks will be displaced, and have to look for careers somewhere else. When retail stores divest themselves of all but one or two humans, when the last manual farm jobs are replaced by recognizing machines, will that be a utopia or a disaster?

[freesoft_2000]

hi everyone,
ai depends on something called fuzzy logic which which are basically comparators with positive and negative feedbacks. This enables the comparator to learn a very simple task thus acting like one brain cell. Thus if you look at the human brain there is the vision part, hearing part, speech part and motion part. The only part that has not been solved is the capability to integrate all these systems to work together thus enabling a robot that can behave and learn like a human.

Yours Sincerely

Richard West

[Chainsaw]

I haven't kept up with research in neural nets, but it seems like most of the research is made in how to simulate a single neuron, and how to feed-back correct answers so they can learn.

But in real brains, it sure seems like the interconnections are the big deal. After all, if we have neurons and cats have neurons, why don't cats learn like we do? It isn't likely that just having MORE neurons makes you smarter. It isn't likely that our neurons are 'better' then theirs. It seems more likely that ours are interconnected better, or that we have neurons devoted to tasks that they don't have, and 'devoted' implies the arrangement, not the quantity.