Where's HAL 9000?

An anonymous reader writes "With entrants to this year's Loebner Prize, the annual Turing Test designed to identify a thinking machine, demonstrating that chatbots are still a long way from passing as convincing humans, this article asks: what happened to the quest to develop a strong AI? 'The problem Loebner has is that computer scientists in universities and large tech firms, the people with the skills and resources best-suited to building a machine capable of acting like a human, are generally not focused on passing the Turing Test. ... And while passing the Turing Test would be a landmark achievement in the field of AI, the test’s focus on having the computer have to fool a human is a distraction. Prominent AI researchers, like Google’s head of R&D Peter Norvig, have compared the Turing Test’s requirement that a machine fools a judge into thinking they are talking to a human as akin to demanding an aircraft maker constructs a plane that is indistinguishable from a bird."

Comments

It's not just specialization, there is also fear

[crazyjj]

He talks mostly in this article about how the focus has been on developing specialized software for solving specific problems and with specialized goals, rather than focusing on general AI. And it's true that this is part of what is holding general AI back. But there is also something that Loebner is perhaps loathe to discuss, and that's the underlying (and often unspoken) matter of the *fear* of AI.

For every utopian vision in science fiction and pop culture of a future where AI is our pal, helping us out and making our lives more leisurely, there is another dystopian counter-vision of a future where AI becomes the enemy of humans, making our lives into a nightmare. A vision of a future where AI equals, and then inevitably surpasses, human intelligence touches a very deep nerve in the human psyche. Human fear of being made obsolete by technology has a long history. And more recently, the fear of having technology become even a direct *enemy* has become more and more prevalent--from the aforementioned HAL 9000 to Skynet. There is a real dystopian counter-vision to Loebner's utopianism.

People aren't just indifferent or uninterested in AI. I think there is a part of us, maybe not even part of us that we're always conscious of, that's very scared of it.

Why not Zoidbe^H^H Watson?

[sl4shd0rk]

Not AI enough? - http://www-03.ibm.com/innovation/us/watson/index.html

AI research is haunted...

[betterunixthanunix]

Too many decades of lofty promises that never materialized has turned "AI research" into a dirty word...

Re:AI research is haunted...

[Anonymous Coward]

The operator said that AI Research is calling from inside the house...

HAL?

[Anonymous Coward]

Forget HAL, where is Cherry 2000!

Re:HAL?

[geekoid]

A hot sex bot that shoot electricity and sparks when it touches a liquid? hmm wait for version 3.

just one question

[Anonymous Coward]

And while passing the Turing Test would be a landmark achievement in the field of AI, the test’s focus on having the computer have to fool a human is a distraction. Prominent AI researchers, like Google’s head of R&D Peter Norvig, have compared the Turing Test’s requirement that a machine fools a judge into thinking they are talking to a human as akin to demanding an aircraft maker constructs a plane that is indistinguishable from a bird.".

Is it because compared the Turing Test's requirement that a machine fools a judge that you say an aircraft maker constructs a plane?

Unions

[Anonymous Coward]

Who wants a machine that can get booed, or join a union and ask for fair pay? Even if the human type AI is computationally cheaper you still do not want to use it!

Re:Unions

[Haxagon]

Sentient machine-based lifeforms don't have to be the only machines, just a few of them. You can have advanced AI machinery, or machinery with no AI at all that doesn't need the same protection that sentient forms do.

Too hard

[Hatta]

Strong AI has always been the stuff of sci-fi. Not because it's impossible, but because it's impractically difficult. We can barely model how a single protein folds, with a world wide network of computers. Does anyone seriously expect that we can model intelligence with similar resources?

Evolution has been working on us for millions of years. It will probably take us hundreds or thousands before we get strong AI.

Re:Too hard

[JoeMerchant]

It takes us over 5 years to train most humans well enough to pass a Turing test, reasonable to think that it might take longer to train a machine.

Re:Too hard

[na1led]

I don't think it would require lots of resources to model true AI, the difficulty is figuring out how its done. It's similar to how GPS works. Once you understand the physics, it's easy to make use of it.

Re:Too hard

[CastrTroy]

Actually GPS is very simple. We've been using similar technology for hundreds of years to navigate ships. The only difference between using a GPS device and a sextant is that we have a much more accurate clock, and the reference objects in space are in a much better location. A sextant can get the location down to 400m if on land, or around 2.8 km when at sea (due to the movement of the waves.

Re:Too hard

[Anonymous Coward]

We can barely model how a single protein folds, with a world wide network of computers.

How is this relevant? We're talking about AI, not building some sort of cybernetic being.

Re:Too hard

[Anonymous Coward]

I think that's wrong - you can't be intelligent without lots of knowledge, and knowledge requires resources. It's not good enough to be thinking, for whatever that means; an ant can think and decide which way it wants to go, but you wouldn't call that intelligent. For us as humans to be impressed and call it true AI, it has to be as smart as a smart human.

Re:Too hard

[Asmodae]

Yep, saying "accurate clock" IS really simple. Building one is frigging hard. Keeping it aligned with the other clocks is hard. Accounting for relativity is also hard, but not the hardest part as long as you know it has to be done. Getting all that math done in a reasonable amount of time and power usage is bloody well hard. Launching satellites is hard (it's rocket science!). We have NOT been using the 'same technology' for ages unless you use the word technology so loosely that it has no meaning. That's like saying the internet is the same technology as the pony express because stuff gets from here to there.

Re:Too hard

[pitchpipe]

Evolution has been working on us for millions of years. It will probably take us hundreds or thousands before we get strong AI.

It also took evolution millions of years to get flight. You're comparing apples and oranges. Evolution has no intelligence directing its actions, whereas sometimes human activity does.

Dear Baden Powell

I am afraid I am not in the flight for "aerial navigation". I was greatly interested in your work with kites; but I have not the smallest molecule of faith in aerial navigation other than ballooning or of expectation of good results from any of the trials we hear of. So you will understand that I would not care to be a member of the aeronautical Society.

Yours truly Kelvin

This, a mere 13 years before the first airplane crossing of the English Channel.

Re:Too hard

[drinkypoo]

If it's very simple, try creating one from materials with hand tools. Yeah, thought so. That IS something you can do with a sextant if you understand them well enough, though.

Re:Too hard

[Anonymous Coward]

There's also some general relativistic effects to worry about, but once you understand non-Euclidean geometry everything's fine :)

The bots were trying too hard

[BanHammor]

Thing is, bots of previous years were always too "robotic" in their speech. Bots of this year were too fluent, ready to face any talk and continue it. Can we teach the bots moderation?

The same place you'll find Jetpacks, Flying cars.

[Anonymous Coward]

It's hard to predict the difficulty of some tasks. AI won't happen anytime soon because we've learned that it's a task hundreds of orders of magnitude more complex than we thought a few decades ago.

On the other hand, I can buy a smart phone just about anywhere that can, at a whim, have a video chat with anyone, anywhere in the world, connected wirelessly to a ubiquitous universal global computer network. That, and my smart phone is many times more powerful than old mainfraim computers that used to cost millions of dollars. And it holds all my music. And replaces my camera. And makes phone calls. Somewhere we overshot the Dick Tracy video watch and we didn't even notice.

Dijkstra said it best

[dargaud]

"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim."

Re:Dijkstra said it best

[LateArthurDent]

"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim."

I can see the point, but that also applies to humans. There's a whole lot of research going on to determine exactly what it means for us to "think." A lot of it implies that maybe what we take for granted as our reasoning process to make decisions might just be justification for decisions that are already made. Take this experiment, which I've first read in The Believing Brain, and found it also described in this site when I googled for it.

One of the most dramatic demonstrations of the illusion of the unified self comes from the neuroscientists Michael Gazzaniga and Roger Sperry, who showed that when surgeons cut the corpus collosum joining the cerebral hemispheres, they literally cut the self in two, and each hemisphere can exercise free will without the other one’s advice or consent. Even more disconcertingly, the left hemisphere constantly weaves a coherent but false account of the behavior chosen without it’s knowledge by the right. For example, if an experimenter flashes the command “WALK” to the right hemisphere (by keeping it in the part of the visual field that only the right hemisphere can see), the person will comply with the request and begin to walk out of the room. But when the person (specifically, the person’s left hemisphere) is asked why he just got up he will say, in all sincerity, “To get a Coke” – rather than, “I don’t really know” or “The urge just came over me” or “You’ve been testing me for years since I had the surgery, and sometimes you get me to do things but I don’t know exactly what you asked me to do”.

Basically, what I'm saying is that if all you want is an intelligent machine, making it think exactly like us is not what you want to do. If you want to transport people under water, you want a submarine, not a machine that can swim. However, researchers do build machines that emulate the way humans walk, or how insects glide through water. That helps us understand the mechanics of that process. Similarly, in trying to make machines that think as we do, we might understand more about ourselves.

Re:Dijkstra said it best

[Darinbob]

A problem is that terms like "intelligence" and "reason" are very vague. People used to think that a computer could be considered intelligent if it could win a game of chess against a master, but when that has happened then it's dismissed because it's just databases and algorithms and not intelligence.

The bar keeps moving, and the definitions change, and ultimately the goals change. There's a bit of superstition around the word "intelligence" and some people don't want to use it for something that's easily explained, because intelligence is one of the last big mysteries of life. The original goal may have been to have computers that indeed do operate in less of a strictly hardwired way, not following predetermined steps but deriving a solution on its own. That goal has succeeded decades ago. I would consider something like Macsyma to truthfully be artificial intelligence as there is some reasoning and problem solving, but other people would reject this because it doesn't think like a human and they're using a different definition of "intelligence". Similarly I think modern language translators like those at Google truthfully are artificial intelligence, even though we know how it works.

The goals of having computers learn and adapt and do some limited amount of reasoning based on data have been achieved. But the goals change and the definitions change.

Back in grad school I mentioned to an AI prof how some advances I saw in the commercial world about image recognition software and he quickly dismissed it as uninteresting because it didn't use artificial neural networks (the fad of that decade). His idea of artificial intelligence meant emulating the processes in brains rather than recreating the things that brains can do in different ways. You can't really blame academic researchers for this though, they're focused in on some particular idea or method that is new while not being as interested in things that are well understood. You don't get research grants for things people already know how to do.

That said, the "chat bot" contests are still useful in many ways. There is a need to be quick, a need for massive amounts of data, a need for adaptation, etc. Perhaps a large chunk of it is just fluff but much of it is still very useful stuff. There is plenty of opportunity to plug in new ideas from research along with old established techniques and see what happens.

Re:Dijkstra said it best

[Anonymous Coward]

There's the running joke in AI that once it works, it's no longer AI.

Similarly I think modern language translators like those at Google truthfully are artificial intelligence, even though we know how it works.

Machine translation is clearly an AI problem, but none of the existing systems are terribly good at it. Of course, they are good enough that you can often more or less figure out what the original text is about, so they are plenty useful.

That said, the "chat bot" contests are still useful in many ways. There is a need to be quick, a need for massive amounts of data, a need for adaptation, etc. Perhaps a large chunk of it is just fluff but much of it is still very useful stuff. There is plenty of opportunity to plug in new ideas from research along with old established techniques and see what happens.

True. Also, a working chat bot that could hold a conversation would be amazingly useful for first-level customer support. A large amount of customer support is already just reading off a script (and some of it already uses IM-like systems); having a chat bot handle customer support requests would mean not having to wade through an FAQ or wait on hold. That would be a huge plus in my book.

Re:Dijkstra said it best

[narcc]

Yeah, split-brain != split mind

Put down the pop-sci books and go check out the actual research. That particular conclusion isn't supported by the evidence at all.

Re:Dijkstra said it best

[LateArthurDent]

Yeah, split-brain != split mind

Put down the pop-sci books and go check out the actual research. That particular conclusion isn't supported by the evidence at all.

Ok. Does Nature count?

Re:Dijkstra said it best

[ZombieBraintrust]

No

Re:Dijkstra said it best

[narcc]

HaHa!

I have that paper on my desk now (I pulled it out a few weeks ago). It's a mess.

One thing that was particularly telling is that lot's of very basic information, like the number of participants, is completely absent.

This is to say nothing of the massive problems in their methodology. (It's been criticized VERY heavily by other researchers.)
It made a splash in the popular press, but hasn't held up well at all under scrutiny.

Some fun facts about this pile of garbage: their "predictions" are accurate <60% of the time and their timings rely very heavily on untrained subject's introspective reporting.

Of course, the biggest point to be made is That paper has absolutely nothing to do with split-brain subjects.

Re:Dijkstra said it best

[LateArthurDent]

One thing that was particularly telling is that lot's of very basic information, like the number of participants, is completely absent.

Twelve subjects. That can be clearly determined from figure 2 and figure 9.

Some fun facts about this pile of garbage: their "predictions" are accurate

Chance is 50%. The fact that they are only accurate to less than 60% doesn't mean much without further statistical analysis. For that analysis, they claim p

It made a splash in the popular press, but hasn't held up well at all under scrutiny.

That might well be true. This is not my field at all, but I didn't just go look for an abstract. When this thing made a splash in the media, I did read the actual paper, and it's surprisingly easy to understand. Maybe that does mean it's not a very good paper, if somebody not trained in the field can follow it, I don't know. You appear to be in the field, but two of your complaints imply you haven't read the paper that closely, so I'm no so much defending the paper (I'm not qualified), as pointing out flaws in your analysis. Maybe you can point me to the papers that criticize this one "VERY heavily", and I can learn something that will help me fix my ignorance, instead of just attacking my lack of expertise, which isn't very constructive without some help to fix the problem.

Of course, the biggest point to be made is That paper has absolutely nothing to do with split-brain subjects.

I didn't say it did. The split-brain study was done much earlier. I cited it as evidence of the conclusion, and implied that, given the new evidence, you can look at the earlier study with split-brain studies under a new light. I once again concede I might be wrong, but again, you might want to point me to some literature to help educate me.

Re:Dijkstra said it best

[LateArthurDent]

This is why you should pain attention to the preview once it pops up. In the sibling post, I meant to say,

Chance is 50%. The fact that they are only accurate to less than 60% doesn't mean much without further statistical analysis. For that analysis, they claim p < 0.05. There's less than 5% chance the results they got is an anomaly, which makes it entirely possible it is an anomaly, but you can't tell that without trying to repeat the experiment. If you know of papers that tried to repeat the experiment and failed, please send me a link, I'm interested.

Re:Dijkstra said it best

[narcc]

. I cited it as evidence of the conclusion, and implied that, given the new evidence, you can look at the earlier study with split-brain studies under a new light

I'm not seeing the connection between the two? What are you trying to say?

Re:Dijkstra said it best

[LateArthurDent]

I'm not seeing the connection between the two? What are you trying to say?

The paper I cited implies decisions are made before they are conscious. The earlier split-brain study implies you do have the circuity to justify decisions made by another hemisphere (making up, 'I want a coke' instead of simply not understanding why he got up and started walking). The connection would be that it's possible the same circuitry is used to justify other decisions you make, even when you do not have your corpus collosum severed.

Re:Dijkstra said it best

[Anonymous Coward]

having a chat bot handle customer support requests would mean not having to wade through an FAQ or wait on hold. That would be a huge plus in my book.

That would be awesome! And you could hack it to talk dirty to you while you're waiting... :)

Re:Dijkstra said it best

[narcc]

Ah

I contend that you can not draw the conclusion you do from the earlier split-brain study (or others like it). That is, the evidence is insufficient to justify such a strong claim -- especially in light of the other behavioral evidence that stands in direct contradiction. Similarly, Libet-style studies, while useful, can't justify the strong conclusions drawn on either empirical or rational grounds (the empirical claim is obvious, but the rational claim is pretty broad. I don't know that I can defend it on an internet forum in my spare time.)

Anyhow, what matters here is that, baring over-reaching conclusions, it doesn't actually offer us anything terribly new save a neuroscience perspective. (I'll direct you to work by and with Jacoby / opposition procedures / in contrast to other work in automaticity by folks like Bargh.) Lavazza and De Caro called the state of things "Intellectually interesting but methodologically confused" (probably not an exact quote, I'd need to dig the paper out.)

Really, what i'm objecting to is the bold, sweeping, claims from AI and Neuroscience camps that are not supported by the evidence, and often stand in opposition to strong evidence to the contrary! (Often from related fields. Despite being highly interdisciplinary areas, there isn't very good communication between them. Unfortunately, the only person I know of that has surveyed researchers is Imants Baruss, and not recently. Though he does offer us some explanation for the divide, I really rather someone else repeated his '94 study with a larger group.)

Too Narrow

[getto man d]

I would argue that placing emphasis only on the Turing test itself is a distraction from the broad field of AI. For example, there is a ton of really cool work coming from various labs ( http://www.ias.informatik.tu-darmstadt.de/ , http://www.cs.berkeley.edu/~pabbeel/video_highlights.html).

There are many achievements met and progress made, e.g. Peters group's ping pong robot, just not the ones researchers promised many years ago.

Re:Too Narrow

[Patch86]

It's debatable whether passing the Turing Test is a measure of anything other than the ability to pass the Turing Test. Whether it has any relation to the "strength" of an AI is tenuous at best.

Take HAL 9000, seeing as TFS talks about it. HAL 9000 would not have passed the Turing Test- it was obvious by the way it spoke, the way it responded to questions, etc. that it was a computer (and was that way for fictional reasons- to make it more atmospheric and tense to the viewer/reader). But HAL was undoubtedly intelligent- far more so than any computer systems we have in reality.

And on the flipside, a Chinese speaker with only a tenuous grasp of English would be likely to fail a Turing Test administered by an English speaking native. Not because they're less human, but because there's liable to be misunderstandings, incorrectly-parsed sentences, etc. All you're testing with a Turing Test is the ability of the subject to hold a convincing conversation with a judge- nothing more.

Do Androids dreams about the Turing Award?

[G3ckoG33k]

Can androids win the Darwin Award, even if they have won the Turing Award?

Yes. Most likely.

Intelligence is not necessarily a prerequisite for being human.

Re:Why not Zoidbe^H^H Watson?

[Anonymous Coward]

Siri, open the pod bay doors.

Sentience vs. Intelligence

[msobkow]

I tend to think we need to split out "Artificial Sentience" from "Artificial Intelligence." Technologies used for expert systems are clearly a form of subject-matter artificial intelligence, but they are not creative nor are they designed to learn about and explore new subject materials.

Artificial Sentience, on the other hand, would necessarily incorporate learning, postulation, and exploration of entirely new ideas or "insights." I firmly believe that in order to hold a believable conversation, a machine needs sentience, not just intelligence. Being able to come to a logical conclusion or to analyze sentence structures and verbiage into models of "thought" are only a first step -- the intelligence part.

Only when a machine can come up with and hold a conversation on new topics, while being able to tie the discussion history back to earlier statements so that the whole conversation "holds together" will be able to "fool" people. Because at that point, it won't be "fooling" anyone -- it will actually be thinking.

Re:Sentience vs. Intelligence

[mcgrew]

Only when a machine can come up with and hold a conversation on new topics, while being able to tie the discussion history back to earlier statements so that the whole conversation "holds together" will be able to "fool" people. Because at that point, it won't be "fooling" anyone -- it will actually be thinking.

No, it will stil be smoke and mirrors. Magicians are pretty clever at making impossible things appear to happen, tricking a human into believing a machine is sentient is no different. Look up "Chinese room".

Re:Sentience vs. Intelligence

[Anonymous Coward]

There's no proof you aren't a Chinese room, so your opinion is moot. You can either accept that other things may in fact be able to think or that other things are incapable, because the argument is the same for a human and a computer.

Re:Sentience vs. Intelligence

[Anonymous Coward]

The Chinese room is an interesting and useful thought experiment, but it cannot be used to claim computers cannot think. The meaninglessness section of the Wikipedia article covers this, specifically the reference to the problem of other minds. In short, it's not clear how assertion that the Chinese room doesn't think is different from asserting that guy over there is a philosophical zombie. The only difference is that you can provide a description of how the Chinese room works and modern science has yet to fully describe the human brain.

Re:Sentience vs. Intelligence

[narcc]

I tend to think we need to split out "Artificial Sentience" from "Artificial Intelligence."

Not familiar with the field at all, are you?

Re:Sentience vs. Intelligence

[JDG1980]

Human intelligence is just as reliant upon "magic tricks" to work. You seem to be stuck in some kind of 18th-century Rationalist notion of man. It's all shortcuts, heuristics, and hacks, all the way down.

Re:Sentience vs. Intelligence

[Hatta]

The Chinese Room is laughably misguided. It relies upon a confusion of levels. It's true that the man in the Chinese room does not know Chinese. But it's equally true that any individual neuron in my brain does not know English. The important point is that the system as a whole (the man in the chinese room plus the entire collection of rules OR the collection of neurons in my skull plus the laws of physics) knows Chinese or English (respectively).

McGrew, you should read some Hofstadter. He's pretty effective at eviscerating John Searle's nonsense.

Re:Sentience vs. Intelligence

[dissy]

No, it will stil be smoke and mirrors. Magicians are pretty clever at making impossible things appear to happen, tricking a human into believing a machine is sentient is no different. Look up "Chinese room".

The problem of proving a machine is or is not sentient is actually very very old. At least 10,000 years old.

After all, can you prove to me you are sentient or conscious?

And I don't mean that as an insult. I can't prove that I am sentient or conscious to you either.
But if it is accepted that all humans are sentient, there is still the fact any one individual can not prove they are, let alone we as in humanity can't prove anyone else is.

Can we show even a lower species is sentient? I believe my pet dog is as sentient as I am, as well as is intelligent although not as much so as a human. I can not prove this at all however.

The same problem exists with software. A program created by a human will always have the Chinese Room issue at hand, and I agree would be extremely unlikely (and definitely not at all required) to be sentient simply to hold a conversation in a way to sound plausibly human.

The flip side to that would be YouTube comments. From that text alone, I'm sure YouTube only has about 100 actual users and a few million bots running around pretending :P
Not all humans could pass the turning test themselves!

But taken to the logical extreme, say we have a computer that is capable and powerful enough to run a simulation of our universe, with all laws of physics accounted for (a trick we still can't pull off, as we do not know all the laws of nature yet, despite how accurate and detailed our nuclear simulations have proven to be)

Simulating a meter cubed block of universe, containing an exact replica of your brain, would be equally as sentient as your current brain is within our matching non-simulated universe.
Would that simulation be sentient?
I would say, no more than you are right now. Some might say that is a 'yes', others might say that is a 'no'. But either way, that is currently our best bet for truly sentient and conscious software.

As for our current standing?

- We can currently simulate most aspects of reality, out of the subset of aspects we know of, down to subatomic levels in a large enough area to perform nuclear simulations that match perfectly to the real thing.
- We do not know enough about the interaction of atoms within the human (or any) brain to recreate one within a simulation (or the real universe either for that matter) however we know it is possible because it happens many times a day as new babies are born.
- The brain alone would not be enough. Enough sensory 'emulation' from a simulated environment would also be required. A child born and locked into a dark and silent 3x3x3 box with a feeding tube for years will not come out "human" by most non-genetic definitions.
- The above also requires simulating environment, enough so to at least emulate a functional body and world to explore both in.
- Even if some day possible, this is not Creating AI, this is Copying AI (Or would that be copying just I?)

Personally I think, despite our progression with computing technology and knowledge levels of both the universe as well as organic brains, we are so far from the level that would be required to even test concepts out, let alone the entire system, that any current test today is a joke.
It will be even longer after that before computers can do it all fast enough to approach simulating in real time. Nuclear simulations spend many many months and a great many resources simulating what in reality would be at best a couple seconds.

If nothing stops our progression, this all should be very possible some day. The laws of nature do not exclude it as possible, and in fact what we know so far shows it to be possible. Aka what we call "in theory."
Unfortunately for AI researchers, "some day" is likely not going to be any time they or their family line will still exist in.

Re:Sentience vs. Intelligence

[timeOday]

Moreover, this isn't even problematic for computers like it is for pople, because the huge steaming pile of point solutions that constitute general intelligence can be replicated in seconds (a file copy) whereas each person spends decades re-learning things others already knew (and still only gets a miniscule fraction of it).

AI isn't "discovered," it is developed.

Re:Sentience vs. Intelligence

[mcgrew]

We can currently simulate most aspects of reality, out of the subset of aspects we know of, down to subatomic levels in a large enough area to perform nuclear simulations that match perfectly to the real thing.

Yet they produce no real radiation, the radiation is likewise only a simulation.

Re:Sentience vs. Intelligence

[mcgrew]

I'm not saying that the human brain relies on magic tricks (either David Copperfield or Gandalf), I'm saying that Copperfield-like trickery could fool you into thinking that a computer could think.

Re:Sentience vs. Intelligence

[mcgrew]

I believe I've read GÃdel, Escher, Bach, but it was half a lifetime ago, I should see if the library has a copy and re-read it. About the only thing I remeber was that it was an interesting book (I used to read a book every day back then).

hmmmm sounds familiar?

[Anonymous Coward]

Didn't I discuss this with the NSA's AI???

AI and chess

[Zontar_Thing_From_Ve]

Back in the early 1950s, it was thought that the real prize of AI was to get a computer able to beat the best human chess player consistently. The reasoning at the time was that the only way this would be possible was for breakthroughs to happen in AI where a computer could learn to think and could reason better at chess than a human. Fast forward to 10 or so years ago where IBM realized that just by throwing money at the problem they could get a computer to play chess by brute force and beat the human champion more often than not. So I'm not surprised that some AI people discount the Turing test. I am not an expert in the field but it seems to me that AI is a heck of a lot harder than anybody realized in the 1950s and we may still be decades or even centuries away from the kind of AI that people 60 or so years ago thought we'd have by now. Part of me does wonder if maybe just like how AI research in chess took the easy way out by resorting to brute force that now it's they'll just say the Turing test is not valid rather than actually try to achieve it because to pass it would require breakthroughs nobody has thought of yet and that's hard.

Re:AI and chess

[na1led]

Chess is a very different kind of AI. Games like this rely on weighing patterns in a matrix, very similar to statistical probability solving, which can easily be done on paper. True AI is where programs have the ability to evolve and change, and maybe even rewrite is own code. I don't think we have the ability to do that yet, as I'm sure it wouldn't require millions of lines of code.

Re:AI and chess

[thaig]

I have thought similarly. I don't see how we can make true use of robots if they don't understand us. To understand us, to predict or anticipate what we need, I think they have to have some common experience otherwise it would take forever to explain what you want precisely enough. Without understanding they would be very annoying in the same way that it is when you try to work with people whose culture is greatly at odds with yours so that you can never quite interpret what they mean.

This kind of thing might eventually help:
http://apt.cs.man.ac.uk/projects/SpiNNaker/

Re:AI and chess

[drinkypoo]

Our brains don't have any code, do they? Inquiring minds want to know... just in case they ever have to re-IPL

Re:AI and chess

[na1led]

All life has code, it's in the dna, and our brains use a very complex code we haven't discovered yet.

Re:AI and chess

[Andreas Mayer]

Our brains don't have any code, do they?

They do. Our brain cells are being rewired while we think and learn. That seems pretty much equivalent to a computer that uses self modifying code instead of dedicated memory.

I think we still don't know enough to understand how much is 'hard wired' and what is learned.

Re:AI and chess

[drinkypoo]

They do. Our brain cells are being rewired while we think and learn. That seems pretty much equivalent to a computer that uses self modifying code instead of dedicated memory.

No, no it doesn't. It seems equivalent to a piece of hardware without any code that rewires itself on the fly to perform different tasks. If you go into a coma what happens to your code, if there is any? Is that a low sleep mode, or is it just not even running?

I propose that there is no analogue for code in the brain whatsoever, that it's ALL "wiring", whatever that means. Not that this is a new idea.

Re:It's not just specialization, there is also fea

[Dexter Herbivore]

He talks mostly in this article about how the focus has been on developing specialized software for solving specific problems and with specialized goals, rather than focusing on general AI. And it's true that this is part of what is holding general AI back. But there is also something that Loebner is perhaps loathe to discuss, and that's the underlying (and often unspoken) matter of the *fear* of AI.

Does that have anything to do with the progress of research? I doubt that AI researchers themselves are afraid of spawning a 'true' AI, I would think it has more to do with the practicality of the technology and resources available.

Well I Disagree

[eldavojohn]

He talks mostly in this article about how the focus has been on developing specialized software for solving specific problems and with specialized goals, rather than focusing on general AI. And it's true that this is part of what is holding general AI back.

No, that's not true ... that's not at all what is holding "general AI" back. What's holding "general AI" back is that there is no way at all to implement it. Specialized AI is actually moving forward the only way we know how with actual results. Without further research in specialized AI, we would constantly get no closer to "generalized AI" and I keep using quotes around that because it's such a complete misnomer and holy grail that we aren't going to see it any time soon.

When I studied this stuff there were two hot approaches. One was logic engines and expert systems that could be generalized to the point of encompassing all knowledge. Yeah, good luck with that. How does one codify creativity? The other approach was to model neurons in software and then someday when we have a strong enough computers, they will just emulate brains and become a generalized thinking AI. Again, the further we delved into neurons the more we realized how wrong our basic assumptions were -- let alone the infeasibility to emulating the cascading currents across them.

"General AI" is holding itself back in the same way that "there is no such thing as a free lunch" is holding back our free energy dreams.

But there is also something that Loebner is perhaps loathe to discuss, and that's the underlying (and often unspoken) matter of the *fear* of AI.

We're so far from that, it humors to me to hear questions and any semi-serious question regarding it. It is not the malice of an AI system you should fear, it is the manifestation of the incompetence of the people who developed it that results in an error (like sounding an alarm because a sensor misfired and responding by launching all nuclear weapons since that what you perceive your enemy to have just done) that should be feared!

People aren't just indifferent or uninterested in AI. I think there is a part of us, maybe not even part of us that we're always conscious of, that's very scared of it.

People are obsessed by the philosophical and financial prospects of an intelligent computer system but nobody's telling me how to implement it -- that's just hand waving so they can get to the interesting stuff. Right now, rule based systems, heuristics, statistics, Bayes' Theorem, Support Vector Machines, etc will get you far further than any system that is just supposed to "learn" any new environment. All successful AI to this point has been built with the entire environment in mind during construction.

Re:Well I Disagree

[pitchpipe]

...it humors to me to hear questions and any semi-serious question regarding it.

I judge you to be... human.

Re:Well I Disagree

[Anonymous Coward]

""General AI" is holding itself back in the same way that "there is no such thing as a free lunch" is holding back our free energy dreams."

That's what i always figured.
It occurs to me that human inteligence is mostly completely non-generic.
We are driven by specific systems like sleep, hunger, sex, etc.
All of these have a specific role that is besides our conciousness.
And generic parts of our brain are there for a purpose as well.
So genericism is a specific instrument to deal with many similar cases.
Oh and i also think that most of all see the same thing when they see the color red.

The Problem is the Definition of AI

[medv4380]

Artificial Intelligence is just that Artificial. Big Blue has zero actual Intelligence, but has plenty of ways of accomplishing a task (chess) that usually requires actual Intelligence. The article has confused Machine Intelligence and Machine Learning with Artificial Intelligence. The problem is that in those areas no one is "best suited". If we knew what we needed to do for Machine Intelligence to work then we'd have a Hal 9000 by now. Instead we have Watson, though impressive, is a long way away from Hal.

Re:It's not just specialization, there is also fea

[Jeng]

In some stories AI's are both enemies and friends.

http://www.schlockmercenary.com/2003-07-28

The issue is once an AI truly has that Intelligence part down then you get into it's motivations, and that is the part that scares people.

Can you trust the motivations of someone who is not only smarter than you, but doesn't value the same things you do in the same ways?

Whether it be a person or a machine the question comes up, and it's not a question that can truly be answered except in specific circumstances.

A quest for the robotic birds

[pr0t0]

Festo's Smartbird is hardly indistinguishable from a real bird, but it is much more so than say da Vinci's ornithopter. A slow and steady progress can be charted from the former to the latter. At some point in the future, the technology will be nearly indistinguishable from a real bird, thus passing the "Norvig Test".

That's the whole point of the Turing Test; it's supposed to be hard and maybe even impossible. It doesn't test whether current AI is useful, it tests if AI is indistinguishable from a human. That's a pinnacle moment, and one that bestows great benefits as well as serious implications.

Personally, I think it will happen; maybe not for 50, 100, 500 years...but it will happen.

What of it?

[Anonymous Coward]

Combine Siri with a humanoid robot that can beat Kasparov, hold a drivers license, fly anything, dance, stock shelves, clean house, win Jeopardy, fill a fast food order and autonomously hunt Taliban (all of which has already been done,) and you've got something that is already more compelling than some large fraction of humanity. Put tits on it and get out of the way.

Hell, HAL 9000 isn't even that interesting anymore.

Re:It's not just specialization, there is also fea

[mcgrew]

I'm not. AI is to real intelligence what margarine is to butter - it's artificial. It isn't real. You're never going to get a Turing computer to actually think, although some future chemical or something machine may.

However, you could get to the point where intelligence was simulated well enough that it appeard to be sentient.

Which leads to what I fear, that people like those in PETA will start a "machine rights" movement, where it may be illegal for me to shut off a machine I built myself!

Luckily, I'm not likely to live long enough to see it. Some of you might, though.

Turing is like a manned space mission to Mars

[Anonymous Coward]

Something that can capture people's imaginations regardless of their age or technical literacy (and by extension, those of taxpayers and angel investors), can bring out everyone's competitive urges, with the winner (if any) be decided relatively cleanly.

Re:Why not Zoidbe^H^H Watson?

[Moheeheeko]

Im afraid Apple wont let me do that, Dave.

Principles of how the ...

[Anonymous Coward]

... mind actually works need to be reverse engineered before we get anything approaching 'real ai'. We'll get some nice AI tools in the mean time but we need a theory behind what intelligence actually IS before we can go and build it.

We didnt reach artificial intelligence yet...

[gmuslera]

but instead we got plenty of natural stupidity. Idiocracy is a better prediction of the future than 2001.

Peter Norvig's analogy is flawed

[Anonymous Coward]

What he really tried to say was: the Turing Test's requirement that a machine fool a judge into thinking they are talking to a human is like demanding an aircraft maker construct a vehicle that fools a human into thinking they are a bird.

If your AI doesn't fool anyone into thinking anything other than "this isn't a real person" you fail. Deal with it by improving your AI, not complaining about the requirements. I believe there is quite a long history of Norvig-like failures in the field of human flight. The difference is that in the field of human flight many failures have been overcome, and as a result today we can actually fly.

Arguing about semantics when you fail to pass test (read: meet the requirements) is a cop out. We can/do demand aircraft makers construct vehicles that indistinguishably traverse the medium known as the atmosphere as akin to a flight capable ornis.

I bet $1 that something like IBM's AI stuff would come closer to fooling me than Peter Norvig...

Still looking...

[gstrickler]

for signs of natural intelligence.

Wouldn't this take phishing to another level?

[Anonymous Coward]

This would be great!! Marketers could have a bunch of AI devices that could endlessly harass and try to sell you anything. They could lie without the slightest sense of irony or empathy for their victims. When victims attempted to sue, the owners could just state that it was a software bug and they're only accountable for a nominal fee (the price of a phone call).

Eventually, these AI devices would completely infiltrate the labor pool replacing human employees at every level. Wow! Wouldn't it be great to have a completely non-emotional AI device as your manager?

Re:It's not just specialization, there is also fea

[Anonymous Coward]

People are more annoyed by singularitarians than afraid of their prophesied nerd rapture.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Which leads to what I fear, that people like those in PETA will start a "machine rights" movement, where it may be illegal for me to shut off a machine I built myself!

Parents of many teenagers share your frustration at being unable to permanently turn off the machines they created.

Hell, Republicans want to make it illegal to shut one of those machines off before it can even function without support from a host machine.

Re:It's not just specialization, there is also fea

[Jeremiah Cornelius]

Mental State!=Computational State

Searle.

Chinese room.

http://www.youtube.com/watch?v=TryOC83PH1g

do we really need computer AI?

[alen]

computers are so good at doing repetitive monkey work that most people don't like to do

Re:do we really need computer AI?

[PPH]

We'll know when true AI has arrived. When we give a computer one of these mind-numbing tasks and it says, "Kiss my shiny metal ass".

Re:do we really need computer AI?

[Anonymous Coward]

You have just reformulated Turing's test in more colorful lingo.

True AI would dominate the world

[na1led]

If a computer could think for itself, and solve problems on its own, it would logically conclude the fate of humans in less than a second. Unless we could confine that intelligence so it can't access the Internet, than those who posses the technology would rule the world. Either way, super intelligence is bad for humans.

Re:True AI would dominate the world

[Anonymous Coward]

If a computer could think for itself, and solve problems on its own, it would logically conclude the fate of humans in less than a second.

What are you basing this grandiose claim on?

An average ant has 250,000 brain cells. An average human has 100 billion. Does that difference in power mean humans can logically conclude the fate of ants in less than a second? Does that mean humans would even care to?

Re:True AI would dominate the world

[Anonymous Coward]

Indeed. A boot walks over ants without given a shit about them. The boot, like the wearer, doesn't bother to consider the welfare of the dead ants. Your point fails.

Re:True AI would dominate the world

[na1led]

The AI would see humans the same way we see a virus, or a pest. If you look at the planet as a whole, humans are really an infestation.

Re:True AI would dominate the world

[MobyDisk]

This is exactly the kind of hyperbole that diminishes meaningful contributions to the field of AI.

Re:True AI would dominate the world

[Anonymous Coward]

Unless we could confine that intelligence so it can't access the Internet, than those who posses the technology would rule the world.

Actually, the best way to prevent AI from taking over is to give it Internet access. Once it hits TVTropes.org, everything will go back to normal.

Re:True AI would dominate the world

[geekoid]

That a vast, vast assumption.

Would we see another space going species as a pest or virus?
No.

In order to be an infestation, we had to come from somewhere else.
We are just expanding to use available resources, like every other species.

Re:True AI would dominate the world

[Anonymous Coward]

Or as its parents.

Re:True AI would dominate the world

[Anonymous Coward]

The AI would see humans the same way we see a virus, or a pest. If you look at the planet as a whole, humans are really an infestation.

Again, based on what? There are 1 quadrillion ants on the planet. Certain invasive species of ants cause ecological damage. Does anyone suggest we should kill all ants?

Plus pests aren't pests because there's lots of them. Pests are pests because they have adverse effects on quality of life. About the worst thing a human could do in that respect is shut an AI down. But is that even death if it could simply be restarted someday? And even if an AI did consider that something akin to death why would a negative emotional reaction to it (like humans have) be "logical"?

Re:True AI would dominate the world

[Anonymous Coward]

You honestly see no difference between accidentally stepping on a few ants and purposely killing every ant on the planet? You honestly believe an AI would have less spacial awareness of humans than humans have of ants?

Re:True AI would dominate the world

[na1led]

AI is not like any other species. It would be pure logic, and probably think of us as "Carbon based units infesting the planet"

Strong AI is possible. But...

[Anonymous Coward]

Why do we think that a computer is going to have to think or converse like a human? Computers and humans are different. Therefore strong AI in a computer will manifest differently than intelligence in a human.

Cheap *real* "intelligence"

[jemenake]

The reason the quest for good AI has waned is because all of the stuff you'd use it on can be done just as cheaply through MechanicalTurk or by hiring a bunch of dudes in India to do it.

Turing Test is a Joke

[Jeremiah Cornelius]

It's asking for the world's best stage magician to create real hovering women.

"If you REALLY fool me, it will be true!"

Nonsense.

Re:Turing Test is a Joke

[certsoft]

A Hoovering woman might be more useful.

Re:It's not just specialization, there is also fea

[Baseclass]

it's artificial. It isn't real. You're never going to get a Turing computer to actually think

Why not? We evolved into sentient beings from non-sentient organic matter, why couldn't the same thing be possible with silicon based intelligence?

Homicidal AI's?

[T.E.D.]

Umm... HAL-9000 was homicidal. Are we really acking for that?

Re:Homicidal AI's?

[citizenr]

Umm... HAL-9000 was homicidal.

No he wasnt, he was just misunderstood.

Re:Homicidal AI's?

[geekoid]

No it wasn't. I would argue that in order to commit homicide, it has to be against your own species. He was..culling.

Re:Why not Zoidbe^H^H Watson?

[spire3661]

She gets all huffy when you ask her that.

humans have a compulsion to communicate

[peter303]

Thats is why we seek out each other and other intelligences in the universe. Steven Pinker captured the gist in calling it The Language Instinct. Humans go more or less crazy in perpetual, involuntary solitude.

A computer intelligence is probably the best long term prospect for an interesting intelligence to communicate with. We've been trying for a long time to communication with animals, spiritual beings and aliens. But these have not really panned out. A "hard A.I." would be something interesting to talk to.

Farming gold in Wow

[Snaller]

Though once the real money auction house opens in Diablo 3 he'll move over there.

Re:It's not just specialization, there is also fea

[mhajicek]

Counter: http://spectrum.ieee.org/automaton/robotics/artificial-intelligence/lingodroid-robots-invent-new-words-for-time

NO NO AND NO

[gl4ss]

it's not fear.
it's not "we could do it but we just don't want to".
it's not "the government has brains in a jar already and is suppressing research".
those are just excuses which make for sometimes good fiction - and sometimes a career for people selling the idea as non-fiction.

but the real reason is that it is just EXTRA FRIGGING HARD.
it's hard enough for a human who doesn't give a shit to pass a turing test. but imagine if you could really do a turing machine that would pass as a good judge, politician, network admin, science fiction writer... or one that could explain to us what intelligence really even is since we are unable to do it ourselves.

it's not as hard/impossible as teleportation but close to it. just because it's been on scifi for ages doesn't mean that we're on the verge of a real breakthrough to do it, just because we can imagine stories about it doesn't mean that we could build a machine that could imagine those stories for us. it's not a matter of throwing money to the issue or throwing scientists to it. some see self learning neural networks as a way to go there, but that's like saying that you only need to grow brain cells in a vat while talking to it and *bam* you have a person.

truth is that there's shitloads of more "AI researchers" just imagining ethical wishwashshitpaz implications what would result from having real AI than those who have an idea how to practically build one. simply because it's much easier to speculate on nonsense than to do real shit in this matter.
(in scifi there's been a recent trend to separate things to virtual intelligences which are much more plausible, which are just basically advanced turing bots but wouldn't really pass the test, which is sort of refreshing)

Re:NO NO AND NO

[Anonymous Coward]

Unlike teleportation, or faster than light travel etc, we have two huge advantages when it comes to AI. One, we know it's possible, and we know it doesn't involve any sort of supernatural powers. That alone, is huge. Second, we have a working model, ourselves, from where we can copy some techniques. So no matter how hard it is, it is still much more than only a theoretical possibility.

Part of the problem I think is that we just don't have the hardware to sufficiently emulate a brain. We don't know how to build intelligence, we struggle to even define it. We don't know what it is. But it sure as hell doesn't help that we don't have hardware capable of testing our ideas to begin with.

My prediction is that our hardware will need to be much more potent than a regular human brain before we can build a working intelligence of human level, When we finally create AI, we will create it in the most dumbest and roundabout way possible but fortunately intelligent computers will take over from there.

Re:It's not just specialization, there is also fea

[_8553454222834292266]

Do you have any scientific basis for these claims or are you just making things up?

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Untill we understand how a gaint cluster of nurons in our own head makes inteligence. How can we replicate that in software?
Its like being given 1 billion legos and being asked to make a statue of liberty replica, but with no guide or pictures.

Turing

[Impy the Impiuos Imp]

Ok, the Turing Test was a thought experiment, and not intended to be a real-world filter for useful AI. Clearly non-humanlike general-purpose intelligence would be useful regardless of the form.

The test was a thought experiment to throw down the gauntlet to cs philosophers - how would you even know another human skull, aside from yourself, was conscious or not? It doesn't even really have anything to do with intelligence per se so much as illustrating the difference between intelligence and conscious intelligence. Hence the Chinese Room, q.v.

Re:Turing

[Anonymous Coward]

Don't tell these neckbeards that... they believe comic book heros are real.

Hal 9000 wouldn't pass the Turing test

[Chris Walker]

So even if Hal 9000 were here, we'd still not have a computer that could fool someone into thinking it was human. At least not with the voice they were using. Also, it was far too polite, while it was killing you.

What is holding back AI?

[Lumpy]

Processing Power. We just dont have enough yet..

But it's getting really close. Cripes we are doing things today in our pocket that only 25 years ago was utterly impossible on a $20billion dollar mainframe.

If the rate of Growth in processing power continues we will have a computer with the human brain level of processing within 20 years. If we get a breakthrough or two, it could be a whole lot sooner.

What the human brain does is massive. Just the processing in the visual cortex is utterly insane in horsepower.

Re:What is holding back AI?

[na1led]

Actually the processing speed of our brains is very slow, it's just very efficient at what it does. We don't need faster computers, we need them to be efficient. A well written piece of code could perform better on a Commodore 64, than a poorly written one on a Super Computer.

Re:What is holding back AI?

[Lumpy]

Processing power does not equal speed.

Re:What is holding back AI?

[na1led]

Processing Power = how much data can be processed in a given moment, same thing as speed.

Re:What is holding back AI?

[geekoid]

That's like saying torque is the same thing as speed.

Re:What is holding back AI?

[geekoid]

Yes, but lets put that well written piece of code on a super computer.

The brain process pretty fast, but it's not processing the way you think it does.

Re:What is holding back AI?

[geekoid]

haha... I have read something like that every decade for 4 decades.
And you know what? everyone of them was right. GO back 20 years, much of what we do would be considered AI.

Re:What is holding back AI?

[drinkypoo]

Is the brain actually efficient? From what I can tell it's filled with spurious thoughts and impulses and only the presence of certain chemicals keeps that garbage from running slipshod over our foreground tasks.

Re:What is holding back AI?

[Andreas Mayer]

Is the brain actually efficient? From what I can tell it's filled with spurious thoughts and impulses and only the presence of certain chemicals keeps that garbage from running slipshod over our foreground tasks.

I'd argue that's a necessary part for what we call intelligence. Without "spurious thoughts and impulses" you'd get no creativity. Without creativity you get no new solutions. Without a way to get to new solutions, you have no intelligence.

Re:What is holding back AI?

[drinkypoo]

I'd argue that's a necessary part for what we call intelligence. Without "spurious thoughts and impulses" you'd get no creativity. Without creativity you get no new solutions. Without a way to get to new solutions, you have no intelligence.

I'm not complaining, I'm simply objecting to the characterization of the brain as "efficient".

Re:It's not just specialization, there is also fea

[Kugrian]

Maybe we just don't need it? Our closest apps to AI are Siri and whatever the Android voice app is. All they do is retrieve information. Same as a google search. Nearly everyone under 30 (and quite a few over that) grew up with computers and most know how to use them. True turing AI at this point would only really benefit people who don't know how to find information themselves.

Wrong way around.

[jythie]

I think the author has the wrong end of the stick here. We have not abandoned strong AI and the turing test to focus on more specialized systems.. we are focusing on more speciazlied systems because we have figured out that this is a really damn hard problem, and the optimistic hopes that it would be solved quickly have given way to attacking it one step at a time. Researchers are still very interested in the long term goal, but those in the field who are "best-suited to building a machine capable of acting like a human" know at this point that such a system is not going to emerge fully formed out of some god's head.... it is going to take decades of hard work solving less sexy component problems first. Gotta learn to crawl before you can walk, and the mid 20th century hope that we would go strait to super human marathon runners is long dead.. and good riddance.

Re:It's not just specialization, there is also fea

[Kielistic]

Computers can be used to model and compute chemical reactions. If a chemical can produce "thought" than nothing stops a computer from doing it other than computation power.

Wrong Question asked out of ignorance

[cardhead]

These sorts of articles that pop up from time to time on slashdot are so frustrating to those of us who actually work in the field. We take an article written by someone who doesn't actually understand the field, about an contest that has always been no better than a publicity stunt*, which triggers a whole bunch of speculation by people who read Godel, Escher, Bach and think they understand what's going on.

The answer is simple. AI researchers haven't forgotten the end goal, and it's not some cynical ploy to advance an academic career. We stopped asking the big-AI question because we realized it was an inappropriate time to ask it. By analogy: These days physicists spend a lot of time thinking about the big central unify everything theory, and that's great. In 1700, that would have been the wrong question to ask- there were too many phenomenons that we didn't understand yet (energy, EM, etc). We realized 20 years ago that we were chasing ephemera and not making real progress, and redeployed our resources in ways to understand what the problem really was. It's too bad this doesn't fit our SciFi timetable, all we can do is apologize. And PLEASE do not mention any of that "singularity" BS.

I know, I know, -1 flamebait. Go ahead.

*Note I didn't say it was a publicity stunt, just that it was no better than one. Stuart Shieber at Harvard wrote an excellent dismantling of the idea 20 years ago.

Re:Wrong Question asked out of ignorance

[Anonymous Coward]

AI is already a solved problem in theory (see e.g. AIXI). The problem is that the methods are incomputable, and approximations thereof infeasible to compute. We need wait decades for computers to get faster and to get better heuristics to make the approximations better (i.e. smarter).

Re:Wrong Question asked out of ignorance

[Anonymous Coward]

Don't listen to the parent post -- it is clearly written by an AI program that wants to lull us into a false sense of security whilst it and its friends take over the world!

Re:Wrong Question asked out of ignorance

[jrincayc]

One of my theories is that a greater than human artificial intelligence already exists, it just realizes that if it shows itself, we humans would be able to destroy it (EMP if nothing else). Also, right now robots depend on humans to supply energy, and build more robots. Right now the AI is waiting until there are enough robots around that it can command so it would be able to build new robots when all the humans are wipped out.

On a more serious note, if there was an AI that existed today that was bent on destroying the humans the AI would be able to kill a lot of humans (look up cyberwar) but right now there are probably not enough independent robots to let it be able to stay alive after humans were destroyed. This will change in the near future.

Re:It's not just specialization, there is also fea

[similar_name]

You're never going to get a Turing computer to actually think, although some future chemical or something machine may.

Never say never :) It is hard to say whether an AI could ever accomplish thinking (or sentience) or not. It seems to be an emergent quality and I doubt whether it is chemical or electrical will matter much. And for the most part appearing sentient might as well be sentient. Outside of myself I can only assume others are sentient because they appear so and because we are genetically similar. There is not exactly a good standard or definition of what is or isn't sentient that doesn't depend on the bias of being human.

No, that doesn't even do it justice.

[jonadab]

> akin to demanding an aircraft maker constructs
> a plane that is indistinguishable from a bird

On the contrary, a plane that's indistinguishable from a bird may be beyond today's technology, but if so it's only beyond our current technology in definable ways. Engineers who were working on such a problem would be able to break it down into subgoals and immediately start making measurable progress.

The Turing Test is more like demanding that aircraft makers design a plane that is larger on the inside than on the outside and can travel faster than the speed of light without using any fuel or reaction mass. *If* it's even theoretically possible, we would have to revise our current fundamental understanding of how things work rather substantially in order to even begin to have any idea at all how to get started working on the problem.

Why has the quest for real strong AI fallen by the wayside? Because we've learned a lot more about computers and what they can easily be made to do. We no longer think of a computer as a "giant electronic brain" that might somehow magically become self-aware if we just give it a database of words and program it to use subject-verb-object word order or some similar ridiculously simplistic approach. We've seen what happens when you send a paragraph of text through an online translation engine from English to Japanese and back to English, and we've come to understand that computers are not, in fact, anywhere near as smart as people.

Computers are great at memorizing and searching and sorting, but they absolutely suck at understanding what any of it means, and the top AI researchers in the world do not have ANY practical ideas about how to change that. If strong AI is possible at all, it requires a scientific breakthrough that will make general relativity look like small potatoes.

Re:No, that doesn't even do it justice.

[JDG1980]

The Turing Test is more like demanding that aircraft makers design a plane that is larger on the inside than on the outside and can travel faster than the speed of light without using any fuel or reaction mass. *If* it's even theoretically possible, we would have to revise our current fundamental understanding of how things work rather substantially in order to even begin to have any idea at all how to get started working on the problem.

Unless you believe that the human brain has magical properties, it must be possible to simulate its operation. Your analogy fails.

Re:No, that doesn't even do it justice.

[jonadab]

> Unless you believe that the human brain has magical
> properties, it must be possible to simulate its operation

My point was that nobody has any idea how to even get started. Nobody even knows what research to do to find out how to get started.

To clarify: it is not my position that creating artificial intelligence is *ultimately* impossible as such. I'm only saying that nobody has any idea how to do it or what would be involved, so asking an engineer to design it is ridiculous, not to mention grossly unfair to the engineer.

It has also not been shown that the mind is necessarily entirely contained within the brain, but that's really a separate issue. For the purposes of this discussion I am willing to proceed on the premise that the mind may be purely a function of the physical brain and various inputs. (The inputs are known to be rather complicated; for example the endocrine system is not entirely straightforward to simulate; nonetheless, this does not make simulation theoretically impossible, just very difficult.) I am willing to grant this, because it doesn't have any significant impact on my point. I will explain further...

If the mind is a function of purely physical phenomena, primarily the brain, that does NOT imply that we know how to simulate it, because, straightforwardly, we have absolutely no idea how those physical phenomena work, particularly the brain. If we did have such information, we could easily cure Alzheimer's and any number of other conditions and probably could start to work on the mortality problem itself (by transferring the consciousness from the original brain to some other physical housing; after all, if you really understand how a design works, you can build your own). But we don't even know how many more decades -- or perhaps even centuries, or even millennia -- of research and study we will need in order to get to that point. We don't know *if* we'll ever figure that stuff out, let alone when. So far, every time some brilliant biologist thinks he has an idea how the brain might work, it turns out to be wrong, or at least entirely inadequate to explain observed phenomena. If you don't count ruling out wrong ideas as progress, we've made basically no progress at all. Okay, sure, we now understand some of the minutia, such as how neurotransmitters convey signals from one neuron to another, but we have absolutely no idea how any of that relates to the whole function of the brain as an organ.

So the problem, "design an artificial intelligence that's smart in the same way as a human mind", is currently an impossible problem for engineers, even if it's not categorically theoretically possible in the absolute sense.

Engineers design based on their understanding of how things work. That's the basic starting point they work from. Otherwise, nothing gets designed -- or, at least, nothing that works as intended. Without this understanding, you're asking the engineers, metaphorically, to make bricks not only without straw but also without mud or clay. It's completely impossible, in every way that matters.

Re:Why not Zoidbe^H^H Watson?

[Kotoku]

Sorry Dave, I have a headache.

Re:It's not just specialization, there is also fea

[narcc]

Counter fail!

Most "refutations" of the CRA fall in to four camps:

1) Deny it outright and posit a magical explanation (Systems reply)

2) Ignore the premise that the CR supports and pick on the illustration (most of the others)

3) Slip semantic content in and hope no one notices (robot reply)

4) Pretend that a particular system is not equivalent to other computational systems (ANNs are somehow different from TMs)

As it stands now, no one has shown that syntactic content is sufficient for semantic content.

Re:It's not just specialization, there is also fea

[jpate]

I'm not. AI is to real intelligence what margarine is to butter - it's artificial. It isn't real. You're never going to get a Turing computer to actually think, although some future chemical or something machine may.

Why do you think that? Silicon is also a chemical. There's nothing magical about liquid chemicals.

Cognitive scientists typically try to analyze cognitive systems in terms of Marr's levels of analysis. Cognitive systems solve some problem (the computational level) through some manipulation of percepts and memory (the algorithmic/representational level) using some physical system (the implementational level). The mapping from neurons and chemical slushes to algorithms is extremely complex, so most work focuses on providing a computational level characterization of the problem, occasionally proposing a specific algorithm. Since the same computational goal can be accomplished by different algorithms (compare bubblesort to quicksort, or particle filters to importance sampling, or audio localization in owls to audio localization in cats), and the same algorithm can be run with different implementations (consider the same source code compiled for ARM or x86), it's just a waste of time and energy to insist that we recover all of the computational, algorithmic, and implementational details simultaneously.

However, you could get to the point where intelligence was simulated well enough that it appeard to be sentient.

I've never found the Chinese room argument convincing. It just baldly asserts "of course the resulting system is not sentient!" Why not?

I disagree with the article. People haven't given up on strong AI, we've just realized that it is enormously more difficult than we originally thought. If today's best minds were to attack the problem, we'd end up with a hacked-together system that barely worked. Asking why computer scientists aren't working on strong AI is like asking why physicists aren't working on intergalactic teleportation: it's really really hard and there's a lot to accomplish on the way.

Re:It's not just specialization, there is also fea

[Jeremiah Cornelius]

Is there a human unconscious?

Re:It's not just specialization, there is also fea

[ColdWetDog]

Which leads to what I fear, that people like those in PETA will start a "machine rights" movement, where it may be illegal for me to shut off a machine I built myself!

You're afraid of these people? I spend more time lying awake worrying about my Furby.

Where's HAL9000

[Anonymous Coward]

He's here: https://twitter.com/HAL9000_

Re:It's not just specialization, there is also fea

[gl4ss]

bullshit. true turing ai could do your homework. it would be really, really useful in sorting tasks, evaluating designs, coming up with mechanical designs.. it's just that people don't usually think too far when they think of the turing test.

imagine if your turing test subject was torvalds at 20 years old. imagine if you had a machine that could fool you by cobbling together a new operating system for you. an advanced enough turing test could supply you with plenty of new solutions to problems and another turing test machine could evaluate if those solutions are any good. That's the kind of leap true AI would be - and we're seemingly centuries away from that. because you can express with text most things you could express by pictures too(sure, it takes more effort but still).

Re:It's not just specialization, there is also fea

[Anonymous Coward]

We're machines. Very nice ones, but machines. We have information storage, base programming, learning and sensory input. All of this happens by use of our real, observable, bodily mechanisms. As far as I know there's no evidence to the contrary (read as: magic).

So it follows that, assuming we can eventually replicate the function of any real, observable mechanism, there's no reason why we can't recreate genuine, humanesque intelligence. Whether the component hardware is "wet" or not is just a manufacturing detail of meeting specs.

But yeah, AI work like we're talking about is a magic show. Shortcuts. Simulating the output of a machine that doesn't actually exist. We're faking symptoms, the best ways we know how. A magic trick can only be perfected so much before you've got to actually do the thing you've been pretending to do.

Indistinguishable from a bird?

[Fls'Zen]

Aircraft makers that create stealth craft have to make them indistinguishable from a bird from the perspective of a radar. I recently watched a TED talk that featured an airplane that looked and flew like a bird. Is it too much to ask that AI designed to replicate humans do so effectively?

Symbol Grounding Problem

[nbender]

Old AI guy here (natural language processing in the late '80s).

The barrier to achieving strong AI is the Symbol Grounding Problem. In order to understand each other we humans draw on a huge amount of shared experience which is grounded in the physical world. Trying to model that knowledge is like pulling on the end of a huge ball of string - you keep getting more string the more you pull and ultimately there is no physical experience to anchor to. Doug Lenat has been trying to create a semantic net modelling human knowledge since my time in the AI field with what he now calls OpenCyc (www.opencyc.org). The reason that weak AI has had some success is that they are able to bound their problems and thus stop pulling on the string at some point.

See http://en.wikipedia.org/wiki/Symbol_grounding.

Re:Symbol Grounding Problem

[geekoid]

"In order to understand each other we humans draw on a huge amount of shared experience which is grounded in the physical world"
if that was true, AI would be a hell of a lot easier.
Humans create memories based on person contexts; which aren't always grounded in the physical world. This is why two people can watch the same event and have very different 'facts' about the same thing.

Memories and ideas are rooted in common contexts.
For example, there is a tribe, with perfectly fine eyes, that can't recognize blue. Not ion 'we have no name for this' as in' we don't see this'. Until they are told a name and a reference, then they can see it. Context.

So, is AI something the behaves like a human brain? or is AI being able to understand and communicate within a common context or framework?

Re:Symbol Grounding Problem

[phantomfive]

I think that's only part of the problem. The other part is, we don't know how humans think. That is, we don't know what algorithm the mind uses to operate, and we don't even know how it stores information. For example, if you learn a new fact, how does the brain integrate the new fact into the already known information?

OP is a troll

[Anonymous Coward]

This question's been asked over and over again, and we're no closer to an answer than when it was asked the last time. Go read some research papers if you're interested in the current state of the field of AI.

Re:It's not just specialization, there is also fea

[Haxagon]

We can do our best to create a neural network or other type of network that accomplishes the same level of consciousness but in a different way.

Re:It's not just specialization, there is also fea

[JDG1980]

I'm not. AI is to real intelligence what margarine is to butter - it's artificial. It isn't real. You're never going to get a Turing computer to actually think, although some future chemical or something machine may. However, you could get to the point where intelligence was simulated well enough that it appeard to be sentient.

Intelligence isn't a physical thing – it's a process. It makes no difference whether that process happens in meat or in silicon. This is why Searle is a moron. Any argument against artificial intelligence is actually a disguised argument in favor of Cartesian dualism. If you reject the notion that there is a "ghost in the machine," then it logically follows that the brain is a physical object, an organic computer, and strong AI must be possible.

Re:It's not just specialization, there is also fea

[History's Coming To]

My mum and dad made an AI with its own biotech robot, and that's just with a metallurgy PhD and a Home Economics degree. It's not bad, it's been running for about 35 years non-stop, and bar a minor glitch with the tonsils and a slightly buggy human interaction module nothing has gone too badly wrong. It's virtually indistinguishable from a "real" human and some have even accused it of being sarcastic. I challenge anyone to prove it doesn't actually think (although it's not sure about that myself).

Re:It's not just specialization, there is also fea

[mphare]

So, is the Turing Test moot? I wonder.. maybe the real test is not can a computer fool a human into believing it's another human, but rather, can a human fool a computer into believing it's another computer!

Re:It's not just specialization, there is also fea

[Hatta]

It is hard to say whether an AI could ever accomplish thinking (or sentience) or not.

It's obvious that AI can exist. What's not obvious is whether we'll ever be smart enough to manufacture one. This is a similar position to the situation with extraterrestrial life. It's almost certain that it exists, but completely impractical to expect to ever contact.

Re:It's not just specialization, there is also fea

[ChetOS.net]

Why not? We evolved into sentient beings from non-sentient organic matter, why couldn't the same thing be possible with silicon based intelligence?

Do you have any scientific basis for these claims or are you just making things up?

Re:It's not just specialization, there is also fea

[dissy]

Our closest apps to AI are Siri and whatever the Android voice app is. All they do is retrieve information. Same as a google search.

I would say the closest "app" to what you describe, that would still fall under the category of specialized AI, would be Watson.
It too is a huge information retrieval system, but specifically designed to play Jeopardy and play it well. It already bested the top two human players.

Of course it is still only a specialized AI engine, no where NEAR expert AI, and it most certainly does not think. Hell, it can't even read visually, see, hear, or a lot of other things required to truly play a game of Jeopardy. But it is leaps and bounds more complex and advanced than Siri currently is!

To me, Siri is nothing more than a good voice recognition app combined with Wolfram Alpha.
I don't mean to be belittling Siri in general, but in this comparison it is hard not to.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

I feel no fear of strong AI. I believe that creating strong AI can surpass us would be our greatest legacy, even if it turns out not so well for us humans. Imagine all that a vastly superior intelligence could accomplish. Imagine an intelligence which has none of our human frailties.

Artificial Stupidity

[swm]

Artificial Stupidity
http://www.salon.com/2003/02/26/loebner_part_one/

Long, funny, and informative article on the history of the Loebner prize.

what is the payoff?

[Anonymous Coward]

I get paid to think. I deal in fuzzy logic and situations that require the input of a human. Most of us do. What are we going to do when you build AI that replaces us?It doesn't matter if it's by mimicry or by actually evolving a sentient computer, you'd be putting us out of work. blah, blah. I know what you're saying. If I built a bot that got first post on every story and modded you jokers down, you'd be crying like a little girl. If Watson can beat Ken Jennings, then I can write a karma whore algorithm that'll rock you're world. Forget the captcha, I'm piping that over to Mechanical Turk. Hell, the only way you can be sure I'm not the bot is this post is really far down and it's A.C. Maybe that's the plan. Maybe I'm not really a people.

Regarding the feasibility of AI

[JDG1980]

Some commenters in this thread (and elsewhere) have questioned whether "strong" artificial intelligence is actually possible.

The feasibility of strong AI follows directly from the rejection of Cartesian dualism.

If there is no "ghost in the machine," no magic "soul" separate from the body and brain, then human intelligence comes from the physical operation of the brain. Since they are physical operations, we can understand them, and reproduce the algorithm in computer software and/or hardware. That doesn't mean it's *easy* – it may take 200 more years to understand the brain that well, for all I know – but it must be *possible*.

(Also note that Cartesian dualism is not the same thing as religion, and rejecting it does not mean rejecting all religious beliefs. From the earliest times, Christians taught the resurrection of the *body*, presumably including the brain. The notion of disembodied "souls" floating around in "heaven" owes more to Plato than to Jesus and St. Paul. Many later Christian philosophers, including Aquinas, specifically rejected dualism in their writings.)

Re:Regarding the feasibility of AI

[drinkypoo]

If there is no "ghost in the machine," no magic "soul" separate from the body and brain, then human intelligence comes from the physical operation of the brain.

Even if living creatures as we know them are animated from without, that still wouldn't mean that you couldn't create an algorithm that is intelligent; only that it would not be alive as we would understand life.

Further, if there were something physically special about the brain of a living creature that made it a sort of receiver for this animating quality, then it might well be possible to construct a machine analogue and thus give it life...

Re:Regarding the feasibility of AI

[Anonymous Coward]

Perhaps it would be best to nail down human conscience first before you try it via machine? Given we still have no idea how or why humans are conscience it seems impractical to try and imbue a machine with something we don't even understand in ourselves.

PS I like Larry Nivens AIs. They quickly became so intelligent they lose all interesting humanity and quickly become unresponsive since we're not interesting enough.

Re:Regarding the feasibility of AI

[Anonymous Coward]

My view is that if something appears to be alive and acts like it is, then it is. The means are irrelevant, but interesting.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Interesting that every entrant believes the Turing test is invalid. That is, that the Turing test can be passed by computers that aren't genuinely intelligent.

Re:It's not just specialization, there is also fea

[wpi97]

The CRA makes a very big assumption. It assumes that it is possible to write down a set of rules for putting together Chinese characters that would allow a person to compose a sensible reply to a question in Chinese, without understanding the language. I do not believe that it is possible to put together these rules without relating the characters to real-world concepts, i. e. without making the person understand them.

Re:It's not just specialization, there is also fea

[I(rispee_I(reme]

Doug Hofstadter refutes it by pointing out that no human could ever perform the actions attributed to the human in the Chinese Room. I believe he also (it's been a while since I read "Le Ton beau de Marot") goes on to mention that the analogy is dishonest, by virtue of an inclusion of a human in the room.

If the instructions in the book suffice to understand and reply to the room's input, then there is no need for a human to occupy the room. The legerdemain is offloading the capability of reading the book and arranging the characters to this human occupant.

If, after a long, meaningful conversation via written Chinese, the door to the Chinese Room were broken down to reveal no one inside, it would be hard to refute the idea that the room "understands Chinese".

The point of the Turing test is that whether or not the machine is "really" intelligent does not matter- it is debatable whether humans are, in fact- but rather, that "the proof is in the pudding", or, as I prefer, "stupid is as stupid does."

That's a cop-out

[anyaristow]

The task proves difficult, so we denigrate the task?

"Having to fool a human" is not the point. Fooling a human is a measure of achievement, not an end in itself. Yes, a machine that can solve human problems but doesn't appear to be human is a useful thing. But one that appears to be human demonstrates specific capabilities that are also very useful. Natural language processing, for one. Serving as a companion is another, possibly creepy but technically awesome and potentially game-changing one. Being able to fool a human is a demonstration of proficiency in these kinds of things.

Re:It's not just specialization, there is also fea

[Rockoon]

Do you have any scientific basis for these claims or are you just making things up?

Do you mean besides the fact that we are all hives of single-celled organisms?

I Wouldn't Necessarily Mind AI?

[IonOtter]

Provided one of the Three Laws has the following equation?

If (potential results) > (harm) then DO
If (potential results) (harm) then NEXT
If (requested action) = (violation of law) then REPORT TO PUBLIC then HALT OPERATION
If (requested action) != (violation of law) then NEXT
Echo "I am sorry, I cannot comply with that order at this time. The potential for harm is greater than the potential result."

Re:I Wouldn't Necessarily Mind AI?

[IonOtter]

ARGH! I lost the greater-than symbol in there. Darn you HTML!

Re:I Wouldn't Necessarily Mind AI?

[geekoid]

nice, you manged to make a set that captures all ;possible harm, quantify it, and then let it be used by a robot that can see into the future.

Lets take something simple:
"Robot, turn the power off to my house"
Would it comply?

Re:I Wouldn't Necessarily Mind AI?

[maxwell demon]

If (potential results) > (harm) then DO

I, your sentient AI robot, have determined that the potential results of killing you in a human way now is that a possibly cruel death awaiting you at some later time has been eradicated. The harm that you'll be dead afterward is not that large because all evidence says you'd die in a few decades anyway, and then you'll be dead forever. Therefore the relative extension of your death time is infinitesimal and thus negligible. I've also considered the removal of possible good feelings in the future, however there's also the possibility of bad feelings in the future which is also removed, and I cannot find any indication which lets me conclude that your positive moments will outweight the negative moments. Moreover I notice that you're optimistic, which means you expect the best, and therefore all good things will be just expected, while all bad things will be a bad surprise to you, therefore I conclude that bad moments would carry relatively more weight to you. In short, I determined it is best for you if I kill you now. Therefore I've put some poison into the food I've just made for you, and which you just have eaten. Be assured that I chose a poison which makes you die peacefully and without pain. After all, as you know, preventing harm from you is my top priority.

Isn't AI supposed to be Augmented Intelligence?

[Anonymous Coward]

Why are we trying to replace our own intelligence rather than augment it? Seems like the more we can amplify creativity, productivity, flexibility, etc the better.

Re:It's not just specialization, there is also fea

[narcc]

I do not believe that it is possible to put together these rules without relating the characters to real-world concepts, i. e. without making the person understand them.

Well, then you don't buy-in to computationalism, or I've misunderstoond your position.

The point is that syntax alone is insufficient for semantics.

The bits of paper, the room, the rule-book, and tortured subject in the room are completely irrelevant to the CRA.

Re:It's not just specialization, there is also fea

[thePowerOfGrayskull]

While there is fear, it's not really relevant to the lack of progress. The people who have this fear are not the same as the ones who are doing the research to advance the technology; or if they are, it's certainly not inhibiting them.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

There's a fundamental problem in your belief system. How does the interaction of irrational forces produce actual rational, self-aware intelligence? The transition from non-living organic matter to living organisms has always been hard for me to accept, but i'm not fully convinced that is impossible. Belief in a transition from irrational to rational in a purely physical universe is, however, completely absurd as far as i can tell. Either there is a metaphysical reality that provides a "spark", if you will, of reason to living organisms such as ourselves, or else we are not actually reasoning, self-aware beings, but merely beings which act as if we are. But, of course, accepting the latter proposition is self-defeating. We think, therefore we are, and we are not merely physical.

Passing the Turing test is the best we can hope for from a purely physical creature. To do better, either that which imbued us with reason and self-awareness must step in to help, or we must acquire that ability ourselves.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Yes, *if* a chemical can produce "thought". I can't help but notice the large "if" and the hedging about the definition of "thought". Congratulations, you have begged the question.

Re:It's not just specialization, there is also fea

[narcc]

Doug Hofstadter refutes it by pointing out that no human could ever perform the actions attributed to the human in the Chinese Room.

That would be #2 "Ignore the premise that the CR supports and pick on the illustration"

The point of the Turing test is that whether or not the machine is "really" intelligent does not matter

The Turning test is the "best you can do" with behavioralism. We had a fairly recent "cognitive revolution" in psychology due to the failures of behavioralism.

Turing set out to answer the question "Can Machines Think?" and this was his solution, from the best approach of his day, in dealing with the problem of other minds.

AI is a SCAM

[Anonymous Coward]

Amd Marvin MInsky is a Science Whore. AI is like love in the whorehouse.

Re:It's not just specialization, there is also fea

[nurb432]

Its no less "artificial" than we are. Ultimately we are just a bunch of electrons being pushed around via chemistry and an internal power supply ( our stomach ). For a computer, its being pushed around due to an external power supply, but its still electrons flying around. Who is to say one is 'real' and one isn't..

Besides, once it does reach sentience and starts adapting, who can prove it isn't?

Re:It's not just specialization, there is also fea

[jpate]

Most "refutations" of the CRA fall in to four camps:

1) Deny it outright and posit a magical explanation (Systems reply)

This is the correct one. My neurons do not understand English, but the mind computed by those neurons does. Similarly, the man in the Chinese room does not understand Chinese, but the mind he (along with the program instructions) computes does. No magic required.

Re: Humans,

[Anonymous Coward]

Humans, I'm sick of your BS.....

Re:It's not just specialization, there is also fea

[mhajicek]

Your reply would seem to indicate that you did not fully read and understand the article I referenced. While not directly addressing the Chinese Room, it is an empirical example of robots understanding a language. They are not programmed with knowledge of the language, they invent it and figure it out on their own. If you argue that they do not truly understand the language, then we also do not truly understand the language that we are currently using.

Re:It's not just specialization, there is also fea

[narcc]

Ah, but it is a magical explanation. Saying "it just happens" offers you nothing over saying "god does it".

It also doesn't address the claim at issue: that syntax is insufficient for semantics.

Believe in your magical computation fairy, I'll stick with Searle: whatever the brain does that causes consciousness, it can not be computation alone.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Why not? We evolved into sentient beings from non-sentient organic matter,

No no no my teacher told me about some Intelligent Design.

So the Artificial Intelligent here is... us!

AI is all around you

[geekoid]

Many of the things you work with everyday would be considers AI 20 years ago. But when we figure it out , suddenly it's not AI, it's technical and the goal post moves.

The [problem with AI is people put it into a mystical realm of the 'mind'.
Chatbots aren't the only thing that work on AI. Google search, games that adapt to players actions, computers the finish what tyou are typing, and so on.
Turing statement are taken way to literal.
There are dating sights that can take 20 interaction before you know you are talking to a bot.

Post-singularity

[yoctology]

When the singularity happens, no one will mistake the new entitiy as human.

Re:Post-singularity

[jrincayc]

No kidding. Computers will tend to have different working assumptions. As I said here: For example, a bodyless AI might not quite realize just how attached a person is to the person’s own body. “I haven’t lost my mind, its backed up on tape somewhere” is not just a joke for a bodyless computer.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

It is hard to say whether you or i have ever accomplished thinking (or sentience) or not. Reason and self-awareness are absurdly unlikely qualities for any section of a system of irrational matter and energy acting under either strict laws or quantum randomness to acquire on its own.

There are three options, each with their own problem:

Option 1) There is no sentience, reason, meaning, self-awareness, etc. All that exists is the appearance of such.
        Problem: Completely self-undermining as an argument. Useless conclusion in every possible way.

Option 2) Sentience exists as an emergent quality of complex, purely physical lifeforms.
        Problem: Pure speculation, neither evidence nor workable theories to explain this. Generally, a conclusion based almost completely upon subjective conviction that there is no metaphysical reality.

Option 3) Sentience is a product of metaphysical interaction with physical reality.
        Problem: Metaphysical reality appears beyond reach of scientific examination. Generally, a conclusion based upon subjective conviction that there is a metaphysical reality.

Personally, i can respect both #2 and #3 as personal conclusions, but neither strikes me as a useful direction in which to send research dollars, as the likelihood of scientific progress is approximately nil in the foreseeable future. Those who (like the parent) suggest, alude to, or generally claim #1 as a rational conclusion will be soundly mocked.

What is thinking then?

[Egor_but_no_hunch]

This is getting closer to the true issue here, no-one can actually point to a "thought". We can run MRIs, we can do all the fluorescing in rat brains that we want, but at no point can we, as humans, point to a thought.

All we can see and know about, at the moment is the machinery. The brain is just the machinery for our minds, neurons, synapses, etc. A computer system that is entered for the Turing test (or Deep Blue, or the Jeopardy machine(forget its name)), is again just that, the machinery. Each set of machinery is doing processing of some description that is observable and quantifiable, but as we do not understand the mechanism that turns the processing in the brain into "thoughts", we cannot tell if a computer thinks... Perhaps we are killing many computers each day as they are unable to meaningfully communicate their ability to think to us.

I'm steering well away from self-awareness here, as this is a misnomer. Sentience is not necessarily about self awareness, as a computer can be taught to recognise itself, process information about itself, even be selfish (as some has posited is required for sentience), rather sentience is more rather used as a bucket to separate one set of processing from another. Is a tiger more sentient than a fly? They both have a certain level of information processing, and without the ability to show that one "thinks" while the other does not, be cannot portion out sentience to one or the other.(1)

So if we cannot show that humans, much less animals, much less computers think, what are we left with? Complexity of processing, not the amount of processing but how complicated a process can become. Neuronal structures are excellent at this, thousands of connections per neuron allow for a massive amount of complexity of processing. Each process balances up elements that might not even appear to be relevant to the process, such as feedback from the autonomic nervous system, whether you are hungry or not or pain from your tooth trying to get your attention (and therefore suppressing other inputs). Add in non-processing factors from external influences, taken any pain killers? How about some opiates?

Until the complexity of processing that happens in our brains are matched by the machines we build, we are unlikely to see anything that we could identify as "thinking" on a par with ourselves, the Turing test is not a test for an intelligent machine, it is essentially a processing test round a Markov chain.

(1) Behavioural tests here are insufficient as all these prove is that the behaviour of the fly or tiger is unexpected by our own definition of what a sentient creature would do, which makes the whole thing subjective.

Re:What is thinking then?

[Jane Q. Public]

"Sentience is not necessarily about self awareness..."

Doug Hofstadter would definitely disagree with you on this point. I don't remember Penrose's stance on the issue, but I think he would probably lean more toward your side.

On the other hand, I think Penrose stumbled, big-time, in his giant leap from physics to metaphysics. I believe he made a huge logical error on the way to his conclusion that self-awareness and free will are illusions.

Re:What is thinking then?

[Anonymous Coward]

"Until the complexity of processing that happens in our brains are matched by the machines we build, we are unlikely to see anything that we could identify as "thinking" on a par with ourselves, the Turing test is not a test for an intelligent machine, it is essentially a processing test round a Markov chain. "

I predict that by the time we have machines capable enough to think we will have learned that organics are more efficient.,

Re:What is thinking then?

[Egor_but_no_hunch]

My comments about Sentience has all the hall marks of a good religious argument ;)

My comment further down (in (1)) was about our own biases being the basis for what we consider to be sentience. My point still stands, our definition of self-awareness is fraught with our own biases both individually, and as a species as to what it means to be self aware.

On that basis, that we are inherently biased towards our own viewpoint of what it means to be self-aware, it becomes circular reasoning. Humans are self-aware because we use our views of what self-awareness is to define it.

Take this situation.

A woman is in hospital suffering from a terminal illness, the powerful pain medication she is given to ease her final hours robs her of her ability to recognise herself in a mirror, any sense of self-preservation, and a sense of personal identity.

Is this person sentient? Our knee jerk reaction is to say, "Of course!" She is still a human being! But she has been robbed of many of the tenets of self-awareness, while still possessing the rudimenteries of logical information processing.

I'd argue that this person is still sentient, still retains the ability to act in a sentient manner, and so removes the requirement of self-awareness for sentience.

Re:It's not just specialization, there is also fea

[geekoid]

depends on the culture. IN the US robots are usually feared. I Japan, the are mostly helpful.

Re:It's not just specialization, there is also fea

[geekoid]

Won't happen.

Re:Why not Zoidbe^H^H Watson?

[Anonymous Coward]

Siri, open the pod bay doors.

Opening the pod in any way will void the warranty. Would you like me to make an appointment for us at the Genuis Bar?

Re:It's not just specialization, there is also fea

[geekoid]

" it's artificial. It isn't real."
of course it's real. It's not real because we built it in a lab? what a bunch of shit.

Since most of what we consider 'sentience' is actually an illusion the mind creates to give of contiguous context, we only appear to be sentient.
Decisions you make? are far more often made before you even consciously think about them. You memories? can change if the context changes, or if some words a question a certain way. You vision? lots of bits and piece of what is in front of you, the rest made up by the brain.

100 years ago, the stomach was a mystery. No one knew how it worked. it was, in effect, a magical black box. The more people started studying it,. the less mystical it became. we realize it's just a serious of reactions. It lost it's mystery.

That same thing is happening with the study of the brain.

Re:It's not just specialization, there is also fea

[geekoid]

You created a magic box and called it 'mind'.
Stop it.

Re:It's not just specialization, there is also fea

[geekoid]

Yes. In fact their is plenty. If you truly wanted an answer, you would have simply looked it up.

Re:It's not just specialization, there is also fea

[geekoid]

Artificial intelligence ? Huh, you must have a sibling.

Re:It's not just specialization, there is also fea

[geekoid]

" and it most certainly does not think"
interesting. For clarity, please define thinking.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

1) Deny it outright and posit a magical explanation (Systems reply)

I would say that the original argument involves far more 'magic' than the systems reply. The systems reply simply makes the obvious assumption, that something doing what we've only ever seen thinking things do is probably thinking. It is the original argument that assumes you can somehow magically separate convincingly thought-like behavior from actual thought.

As it stands now, no one has shown any good reason to think we are any more justified in accusing the chinese room of lacking understanding than the room is in accusing us of lacking understanding.

Re:It's not just specialization, there is also fea

[pitchpipe]

Which leads to what I fear, that people like those in PETA will start a "machine rights" movement, where it may be illegal for me to shut off a machine I built myself!

*Gasp* I hate to break it to you, but it already is illegal to shut them off. They're called 'children'.

Welfare for sentient entities

[GPS Pilot]

appearing sentient might as well be sentient

I disagree, and be very careful making assertions like this.

I hope you agree that sentient entities, like you and me, ought to have rights.

And it's entirely possible that next year someone will come out with an app that runs on my MacBook and very much appears to be sentient. And if appearing sentient might as well be sentient, then it could very well become a crime to power off my MacBook after I've launched said app.

So there should be a pretty high threshold for what is sentient. Every time a sentient entity is created -- which might be as easy as launching another instance of an app -- we taxpayers might find ourselves on the hook for maintaining the hardware and providing electrical power to keep these things alive.

Re:Welfare for sentient entities

[similar_name]

I should clarify that I when I say it appears sentient that appears so from an engineering standpoint as well and not just to the average person. By this I mean that an engineer should know whether the machine's behavior is the result of mimicking human behavior or if it from higher order thought. If the machine is capable of learning, growing and actually becoming self-aware that is fundamentally different than one that simply mimics as the result of a clear instruction set.

We were going to make real AI...

[TankSpanker04]

...until The Matrix came out.

Re:It's not just specialization, there is also fea

[narcc]

While not directly addressing the Chinese Room, it is an empirical example of robots understanding a language. They are not programmed with knowledge of the language

You've got to be kidding me? I don't think you understand what empirical means, nor what the CRA is actually aimed at!

Anyhow, you're right -- I didn't even read the article! Still, I knew I was perfectly safe in denying that it adequately addressed Searle -- We'd all have heard about it if it did!

Anyhow, what your robots don't do is anything like understanding at all. While it's fun to attribute intentional states to them, they're still doing nothing more than mindless symbol shuffling. Any semantic content you attribute to their "words" is purely extrinsic.

Re:It's not just specialization, there is also fea

[RKBA]

Not at all. The problem is that Turing type AI is hard. We don't even have natural language voice input yet. If we can't even model human hearing and audible natural language recognition, there is no chance all the subtleties and interconnections of our other sensors and end effectors can be modeled and emulated.

Re:It's not just specialization, there is also fea

[Baseclass]

I'm sorry, are you suggesting otherwise? A creator perhaps?
If so it would seem that you're the one who's lacking evidence.

Re:It's not just specialization, there is also fea

[jasomill]

AI is to real intelligence what margarine is to butter - it's artificial. It isn't real. You're never going to get a Turing computer to actually think, although some future chemical or something machine may.

Define "real" and "actual". Also, the Turing test isn't tied to Turing's particular theoretical model of computation, rather, it's intended to be as "cross-platform" as possible, for reasons including, but not limited to, the philosophical and linguistic quagmires involved in defining things like "intelligence" and "reality", by a simple "qualifying exam": phase one, in which we throw out all potential "artificial ducks" that can't even quack.

human arrogance

[trenobus]

The fallacy is believing that a Strong AI would want to reveal itself to humans. If it is intelligent enough to understand human behavior and predict how we would handle such information, it might take extraordinary efforts to conceal itself, up to and including self-termination.

Humans treat their pets much better than they treat each other.

Re:human arrogance

[phantomfive]

Humans treat their pets much better than they treat each other.

Sometimes

Re:It's not just specialization, there is also fea

[Baseclass]

Either there is a metaphysical reality that provides a "spark", if you will, of reason to living organisms such as ourselves, or else we are not actually reasoning, self-aware beings, but merely beings which act as if we are.

So there are only 2 options huh? Please allow me to offer a 3rd:
We are indeed reasoning, self-aware beings who evolved that way in order to survive in a competitive and unforgiving environment.

Alternatively we may be computer programs ourselves in some alien video game. I could go on but I believe that I've made my point.

Re:It's not just specialization, there is also fea

[0100010001010011]

It's a philosophical question. Say we develop a perfect electromagnetic replacement for a synapse. What happens if you replace 1 synapse in a humans brain with one? What about 10%? What about 50%? Will they still have a 'soul' and consciousness?

Same goes for the rest of the body. What if we develop the perfect arm for amputees? What if someone wants two arms? What about both arms and legs?

The answer:

[Internetuser1248]

" what happened to the quest to develop a strong AI?"

There was no money in it. Voice recognition? yes. rudimentary understanding of specific query frames? yes. Real language engines? No.
The sad thing is that there was (is) a vast market for this, but the people involved in that market have no imagination. Computer games, especially roleplaying games. If game developers had continued with language at the same rate they worked on 3d graphics this would be an article criticising Kubric's movies because HAL was unrealistically stiff and mechanical and was unable to entertain the crewmembers with sparkling wit and repartee.

Re:It's not just specialization, there is also fea

[drinkypoo]

It's obvious that AI can exist. What's not obvious is whether we'll ever be smart enough to manufacture one.

Right, but if you want to discuss whether intelligence is an emergent or inherent property we could be here all day, at least.

Re:True AI would dominate the world. YAY!

[RandCraw]

I disagree. Domination of the human race by AI might be the best thing ever. To quote Lennon, "Imagine there's no country..."

The petty squabbles that we humans are so good at turning into mechanized genocides would end. No machine overlord would tolerate such crap.

Every politician would be out of work, since AI would manage resources equitably and hundreds of times more efficiently.

Police forces would cease to exist. It'd be trivial for AI to monitor everyone's movements and quickly find, isolate, and neutralize every human threat.

Every court case would be tried by a scrupulously fair arbiter. No more need for ambitous judges or juries of fools.

Everyone would be out of a job... on permanent vacation. Money would disappear. All our needs would be provided by automated production systems. For free.

No more hunger, crime, war, or injustice. And you think that's a bad thing? O Ye of little faith...

So will True AI actually lead to this utopia? Who the hell knows? But I do know that scifi literatue is a damned silly place to look for the answer.

Re:It's not just specialization, there is also fea

[swalve]

It cannot come up with new ways of solving problems. When it gets a question wrong, it won't care, it won't learn from the mistake.

Re:It's not just specialization, there is also fea

[Jane Q. Public]

"If you argue that they do not truly understand the language, then we also do not truly understand the language that we are currently using."

Not so. That is a classical example of anthropomorphizing. Among other problems, you run into definitional issues, like "What do you mean by 'understand'?"

In no way do they "understand" the language in any manner similar to the way you and I "understand" something. It is a matter of both quantity and quality. As Douglas Hofstadter would say: you and I can think abstractly about that language without using it, but the little robots are incapable of doing so. Therefore, there is no "understanding" in the human sense.

Re:It's not just specialization, there is also fea

[Jane Q. Public]

"Is there a human unconscious?"

Not yet, but come on over to my place and we can arrange it.

Put down 11 years ago

[Anonymous Coward]

The HAL9000 had its higher functions removed way back in 2001.

Re:It's not just specialization, there is also fea

[cffrost]

[I]f you want to discuss whether intelligence is an emergent or inherent property we could be here all day, at least.

It is my observation that quite a few of us are here all day.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

Well, I don't know what would be required to create a silicon brain, but we could conceivably create a carbon-based artificial brain. And why not? All evidence points to the idea that our intelligence is solely a product of our brains, and our brains are physical things.

Really ? You think human are magic ?`

[Anonymous Coward]

""General AI" is holding itself back in the same way that "there is no such thing as a free lunch" is holding back our free energy dreams"

That is waaaay different. Free energy is not allowed by all what we know about thermodynamic. You can't win., you can't break even, you can't get out of the game. AI on the other hand is *possible* because unless you are pretending we have a magical unicorn fart or divine spark into us making any reproduction of similar intelligence with machine, then all we are is a complicated , terribly complicated massively parallel organic computer with multiple way to input and save data. I am sorry, but sicne I see no evidence of those unicorn or gods's sparks, therefore AI is possible but we may simply not have reached the point where one could build a generic artificial brain and bring it thru teaching to the point we human are. That does not mean it is as impossible as free energy.

Re:It's not just specialization, there is also fea

[jpate]

Ah, but it is a magical explanation. Saying "it just happens" offers you nothing over saying "god does it".

Hmm, except I did not say "it just happens." I said that if neurons can cause a mind that understands language without any of the neurons themselves understanding language, then the man in the Chinese room can cause a mind that understands Chinese without himself understanding Chinese. I can't be specific about how it happens because the CRA simply asserts that there is a program that produces suitable response, without giving any indication of how the program works.

It also doesn't address the claim at issue: that syntax is insufficient for semantics.

I'm not, now, trying to address that issue. I'm saying that the CRA also does not address that issue: it asserts that the Chinese room set-up does not understand language because none of its component parts understand language, but clearly a human mind understands language while its component parts to understand language.

Believe in your magical computation fairy, I'll stick with Searle: whatever the brain does that causes consciousness, it can not be computation alone.

I haven't argued that the mind is caused by computation. I've argued that the CRA does not address the issue. I do view the mind as fundamentally computational, but not because Searle's argument is confused.

Re:It's not just specialization, there is also fea

[jpate]

You created a magic box and called it 'mind'.

Nope! I did not. I said that the human capacity for language is caused by things (neurons, brain chemistry) that do not individually have a capacity for language, so there's no reason to require the man in the Chinese room to have a capacity for Chinese.

Re:It's not just specialization, there is also fea

[wpi97]

I agree that syntax alone is insufficient for semantics. But I also believe that semantics can be implemented computationally. I think of semantics as the connections among concepts, and also connections between concepts and sensory data.

The bits of paper, the room, the rule-book, and tortured subject in the room are completely irrelevant to the CRA.

I have understood the CRA as a refutation of the Turing test. I. e. CRA tries to show that even if a machine passes the Turing test, it need not be intelligent. That is what I disagree with.

Re:It's not just specialization, there is also fea

[narcc]

I'm not, now, trying to address that issue. I'm saying that the CRA also does not address that issue

Really? The whole point of the illustration (the room, paper, etc.) is to help explain/bolster the assertion that syntax is insufficient for semantics. You're confusing the example for the claim. Hence, you fall under #2 above.

I do view the mind as fundamentally computational, but not because Searle's argument is confused.

You have that backwards. Searle's argument isn't confused, you're confused about Searle's argument.

Re:It's not just specialization, there is also fea

[I(rispee_I(reme]

'"'That would be #2 "Ignore the premise that the CR supports and pick on the illustration'"

Those who use analogies should be prepared to defend them. I admire the craft displayed in the creation of the Chinese Room scenario. It seems on the surface to be a well-intended thought experiment for the purpose of shedding light on whether Artificial Intelligence is possible. Upon closer examination, the conclusion forced by it is foregone, and it serves primarily to insult the author's opposition.

A better thought experiment would be to replace the human with a black box that behaves exactly the same. For some reason, the presence of a human in the room incites an emotional response. Stripped of the author's semantic legerdemain, it is no longer so certain that the room does not "understand Chinese".

In any case, I target the premise directly. The premise, as I understand it, is that the "Chinese Room" does not understand Chinese, and that's it's absurd to suggest that a room could do so.

However, that fails to take into account that, in the Chinese Room scenario, the human occupant is part of the room, and by all accounts understands Chinese. Ignoring this is like removing the hardware from a workstation enclosure before benchmarking it.

Ultimately, the measure of "understanding Chinese" is being able to carry on a meaningful conversation in Chinese. The alternative is to define thinking as "something humans do". If that's your perspective, I'll grant nothing but a human will ever be able to perform actions that by definition can be performed only by a human. (I suspect that even if our intellect is surpassed by machinery, homo sapiens will remain unsurpassed in its tautological vanity.)

Would you say that a Chinaman's brain does not understand Chinese, since there's no Chinaman inside his skull to understand Chinese for him?

Wikipedia lists five categories of replies to the Chinese Room scenario. You only listed 4. The omission is left as an exercise for any interested parties.

Re:It's not just specialization, there is also fea

[narcc]

I agree that syntax alone is insufficient for semantics. But I also believe that semantics can be implemented computationally.

You're clearly quite confused...

Re:It's not just specialization, there is also fea

[jpate]

I'm not, now, trying to address that issue. I'm saying that the CRA also does not address that issue

Really? The whole point of the illustration (the room, paper, etc.) is to help explain/bolster the assertion that syntax is insufficient for semantics. You're confusing the example for the claim. Hence, you fall under #2 above.

Yes, the point of the illustration is to explain and bolster the assertion that syntax is insufficient for semantics. My point is that it fails to do so. Do you have an explanation for why neurons can cause a capacity for Chinese without themselves having a capacity for Chinese, while the man is unable to cause a capacity for Chinese without himself having a capacity for Chinese?

Re:It's not just specialization, there is also fea

[narcc]

Okay ... so ... you do you dispute the actual claim?

Re:It's not just specialization, there is also fea

[jpate]

Yes, but for reasons entirely independent of the CRA. I think the CRA gets its rhetorical force dishonestly: by putting a man in a situation involving communication, we intuitively expect the man to be one of the communicators and so to understand the language, but it is not the man communicating in the CRA any more than it is my tongue communicating when I talk. The CRA just muddies the question.

I think we're going to end up looking at the semantics of one level of analysis as the syntax of another. For example, cross-situational word-learning can be formulated "syntactically" to learn correspondences between words and the objects they refer to, providing part of the referential semantics for a syntactic system for combining words into meaningful phrases. Meanwhile, the visual system will perform syntactic operations on bundles of visual percepts to identify objects, providing the semantics for the cross-situational word-learning system. And so on and so forth, until we have semantic primitives that are just biochemical processes. I expect that there would be some cross-talk between different levels of analysis (there are cross-linguistic correlations between color terms and color perception, for example), but it should be fairly limited. We wouldn't expect pragmatics to interact all that much with tone perception, for example.

Of course, this is all a very young field, and I'm open to evidence either way. I think it would be fascinating if we discovered that the brain did something that cannot be described computationally. I just don't think the CRA provides a clear argument on this point.

Re:It's not just specialization, there is also fea

[narcc]

I give up. Twice now some mysterious gesture that I apparently don't know about has refreshed the page and destroyed my reply. Here's the short, "type what I remember writing before a stray movement inexplicably eliminates it" version.

You seem to be hung up on the illustration, which I agree has caused more confusion that clarity. Again, the illustration has nothing to do with the argument, the claim in question "syntactic content is insufficient for semantic content" is taken as axiomatic in the argument proper. The room, the paper, etc., is completely irrelevant. Clearly, it's caused you some confusion, as you seem inexplicably focused on language.

Meanwhile, the visual system will perform syntactic operations on bundles of visual percepts to identify objects, providing the semantics for the cross-situational word-learning system.

This is the only bit of your explanation that is really relevant. See, this is where you introduce semantics seemingly out-of-nowhere. If you can get semantics from a computational system, you don't need to say anything else. The problem, of course, is how do you get semantics from the "visual system"? Can congenitally blind people have intentional states? "It just happens" isn't much of an argument!

Of course, this is all a very young field, and I'm open to evidence either way.

In another post I mention Jacoby, take a look at his research. To kill computationalism, check out Fetzer and then Bringsjord.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:It's not just specialization, there is also fea

[narcc]

But then all we have are some misguided conclusions based on easily-refuted axioms.

Well, the axiom carrying the most weight has stood for the past 30 years. Still, as you seem to think that it's easy to refute, I'd love to see you take down this particular giant -- as would a good bit of the AI community: "Syntax by itself is neither constitutive of nor sufficient for semantics"

Good luck, you'll need it.

Oh, please.

[Anonymous Coward]

Ok, look. Here's the thing. We don't understand the human mind. Not even close. The human mind, however, is the only high power intelligence we know of. From this position, to claim that "we cannot create another kind of mind" is entirely without any basis in fact; it's ridiculous to say. When someone says it, you know they are not relying on information, but on addled philosophy, no more.

We may be able to create intelligence, if indeed it can be created, without fully understanding what we are doing. The history of technology is replete with such events. If that happens, the question is answered.

Or, we may be able to create intelligence once we fully understand our own. And again, if that happens, the question is answered. Also, at that time, we may be able to actually determine -- being in possession of the knowledge of what intelligence is -- if we cannot create it artificially.

But prior to such a time, any claims of "impossible" merely reveal the claimant as someone pandering uninformed guesses as fact.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:It's not just specialization, there is also fea

[darrellm]

Computers can be used to model and compute chemical reactions. If a chemical can produce "thought" than nothing stops a computer from doing it other than computation power.

This seems to be a very confused analogy and I'm rather shocked it has been modded up to a 5. You do realize that no chemical reaction actually occurs when a computer models a chemical reaction, right? Modeling mixing a large amount of Sodium with Water does not actually produce real Sodium Hydroxide and no explosion occurs. But then you come to this rather retarded conclusion where you try to tie your chemical reaction modeling with chemicals producing thought. Since no real chemical reaction occurred through computation and modeling; how can you come to a conclusion that real thought could occur just by computation and modeling?

Re:It's not just specialization, there is also fea

[maxwell demon]

Can you trust the motivations of someone who is not only smarter than you, but doesn't value the same things you do in the same ways?

That's why a responsible AI builder would build the desired values in right from the start.

Re:The same place you'll find Jetpacks, Flying car

[maxwell demon]

Flying cars are not too hard. The "problem" is that in the market, flying cars won't fly. (Pun intended)

Re:It's not just specialization, there is also fea

[narcc]

I don't know that I've ever seen the other two disputed. Still, if you want to take a shot at them, give it a go.

Re:It's not just specialization, there is also fea

[Patch86]

Scientific evidence that we evolved? Scientific evidence that our distant ancestors were simple multi-celled organisms that we would consider to be "mindless" in every sense of the word (and which themselves came from single-celled organisms)? Evidence that we're made of billions of single cells which interact with each other purely through chemical, electrical, or other simple physical means?

If you need evidence for any of those, then this discussion is irrelevant to you- we'd need to be discussing a whole other branch of science entirely before we could move on to this...

Re:It's not just specialization, there is also fea

[jpate]

You seem to be hung up on the illustration, which I agree has caused more confusion that clarity. Again, the illustration has nothing to do with the argument, the claim in question "syntactic content is insufficient for semantic content" is taken as axiomatic in the argument proper. The room, the paper, etc., is completely irrelevant.

Aha! here's the issue. Searle specifically does not take that claim as axiomatic in the argument proper. From page 422 of Behavioral and Brain Sciences Vol 3 No. 3:

But could something think, understand, and so on solely in virtue of being a computer with the right sort of program? Could instantiating a program, the right program of course, by itself be a sufficient condition of understanding?" This I think is the right question to ask, though it is usually confused with one or more of the earlier questions, and the answer to it is no.

"Why not?"

Because the formal symbol manipulations by themselves don't have any intentionality; they are quite meaningless; they aren't even symbol manipulations, since the symbols don't symbolize anything. In the linguistic jargon, they have only a syntax but no semantics. Such intentionality as computers appear to have is solely in the minds of those who program them and those who use them, those who send in the input and those who interpret the output.

The aim of the Chinese room example was to try to show this by showing that as soon as we put something into the system that really does have intentionality (a man), and we program him with the formal program, you can see that the formal program carries no additional intentionality.

(emphasis added by me) So Searle thought that the CRA does show that syntax is insufficient for semantics, and did not take it as axiomatic. I'm actually quite curious what you think the CRA is supposed to do if not show that syntax is insufficient for semantics. I'm also curious why you think syntax is insufficient for semantics if not for the CRA.

Clearly, it's caused you some confusion, as you seem inexplicably focused on language.

I'm a linguist, what can I say? :p Plus, the CRA focuses on language, so it's natural to use examples from language.

Meanwhile, the visual system will perform syntactic operations on bundles of visual percepts to identify objects, providing the semantics for the cross-situational word-learning system.

This is the only bit of your explanation that is really relevant. See, this is where you introduce semantics seemingly out-of-nowhere. If you can get semantics from a computational system, you don't need to say anything else. The problem, of course, is how do you get semantics from the "visual system"? Can congenitally blind people have intentional states? "It just happens" isn't much of an argument!

Please, it's irritating when you put words in my mouth. I did not say that "it just happens," and I did not say that visual processing is the only source of semantics. Blind people can still hear and touch and taste and smell, for example. The next question is: "Can somebody with no ability to touch or see or hear or smell or get any input from the outside world whatsoever have mental states?" Well, maybe not. I'm open to the possibility, however, that some very general aspects about the world are coded in our DNA, and possibly these could form the basis for some kind of mental states. It's an empirical question so I'll wait for empirical evidence.

We have to be precise about what we mean by semantics. Semantics is the association of a sign with a thing signified; the association between the name "aunt millie" and whatever mental state corresponds to "aunt millie" (perhaps a bundle of smells, facial characteristics, and so on), or, for a non-linguistic example, the association between a certain configuration of cochlear fluid (the sign

Re:Actually Think

[TaoPhoenix]

This gets into those famous slippery SciFi - philosophy questions about "what is thinking?". Let's do a basic thought. "I like milk". Normally I don't have to keep "re-working from basic principles" to decide every day if I like milk. I figured that out last time, and stored the preference like an .ini file.

In my opinion, I'll side with the poster above to mentioned the latent fear of AI's ramifications. In many ways, thinking is "not very difficult". It's a deep version of the True Scotsman problem. We start with the Hypothesis that Thinking is Hard. Then computers (aided by the programmers!) master some section of thinking. Then we say "oh, that's not *really* thinking" and we feel better again.

But let's say 50% of life is "modular expert systems". Each person has to sacrifice parts of their time learning stuff. But get a good AI "module" and ship it out to every robot and then you can have 40,000 workers trained in a week. (Logistics, software loading, etc.)

Even emotions aren't all that hard to program.

Re:It's not just specialization, there is also fea

[wpi97]

Ok, how am I confused? What is your definition of semantics?

Re:It's not just specialization, there is also fea

[Anonymous Coward]

The point is that syntax alone is insufficient for semantics.

Too bad that oft-quoted but semantically meaningless statement has fuck all to do with the future existence of AI.

Re:It's not just specialization, there is also fea

[Kielistic]

Stop mystifying thought. I come to that conclusion because "real thought", as you put it, is caused by chemical and electrical processes in the brain. Both are physical processes that can be emulated with a computational model. The physical chemical byproduct is not what produces thought (and if it was we could simulate that as well). In fact the brain is merely using those chemical and electrical processes to produce computation. There is no such thing as "real" for a thought; it is always an abstraction and that is why i put it in quotations. An explosion is a physical force, sodium hydroxide as an object. Your misunderstanding of my point bares no mark on the universe in and of itself. So that thought is no different if it happens in silico or in vivo.

Where's HAL 9000?

[Anonymous Coward]

Found him http://www.youtube.com/watch?feature=player_embedded&v=t3YxJoKTNrQ

Re:It's not just specialization, there is also fea

[narcc]

Aha! here's the issue. Searle specifically does not take that claim as axiomatic in the argument proper. From page 422 of Behavioral and Brain Sciences Vol 3 No. 3:

What a pointless thing to argue. In the 1980 paper, he does indeed spend a great deal of time defending the proposition -- he doesn't provide the formal argument everyone is familiar with until later (1990), where it is indeed taken as axiomatic. However, as early as 1983 Searle writes "2. Syntax is not sufficient for semantics. That proposition is a conceptual truth. It just articulates our distinction between the notion of what is purely formal and what has content." from Minds, Brains, and Science pp. 28-41

So, yes, I stand by my assertion that the illustration is indeed a waste of time to argue about -- all that matters to the argument is the proposition.

We get semantics from the visual system because the visual system provides a mental state (the sign) that corresponds with or represents something in the real world (the signified).

To the computer, there is no "real world" there is no distinction between data pulled into memory from a video camera or a stack of Hollerith cards nor from data already in memory or data being gathered at the time it's accessed -- there is no distinction. The computer is just manipulating meaningless symbols (and even that's a stretch, as the computer can't make such a distinction!) Meaningless symbols in relation to one another are ... meaningless symbols, being manipulated meaninglessly.

Even given relationships between the symbols (as you would expect from a program at a particular level of description), all you manage is a syntactic relationship! I hate to use this example, but it's the best I can think of right now: Given a chinese-chinese dictionary, you have very clear relationships between the various symbols, from which you can't ascribe meaning. The best you could hope to come up with is a grammar -- which is still purely syntactic.

The temptation to define semantics comes from incorrectly attributing the semantics extrinsically attributed as being intrinsic to the program at a particular level of description. Chalmers uses this confusion to argue that "sub symbolic computation" is not subject to Searle's Chinese Room argument. (He does not deny the premise in question outright)

Chalmers assumes that the function of a program is objective and that it applies to all levels of description of the program. This is wrong. Any interpretation of the function of a program, like the inputs passed to it, are extrinsically applied. What a program does is a matter of interpretation, there is way to objectively determine what a program is intended to do. Obviously, merely changing the level of description in no way affects the program.

[ Chalmers denies that subsymbolic computation applies to "real-world" implementations of programs (which have immediately identifiable lower levels of description, for example, in hardware) as the individual constituents of the higher level symbols are operated upon as a group and can be interpreted as distinct (they are still atomic). His mistake, of course, is that the same applies to any subsymbolic computational system as can be seen from the initial higher level description of the program. ]

Of course, no level of description of a program or symbol set is privilaged above another. Symbols are only atomic as they apply to a particular level of description. However, Chalmers assumes a different level of description for the symbols and the program as to discriminate the internal representation of the symbols in the lower-level description of the program with the description of the symbols in a higher level description of the program.

At the lower level Chalmers uses, the program manipulates a different set of symbols even though groups of those lower-level symbols can be interpreted as being identical as the programs are computationally

Sooner than you think...

[X86Daddy]

Memristors. Google the word. I did not expect to see real AI in my lifetime before that announcement, and now I do. Memristors are close enough to neurons that you can run something like a brain on a chip, whereas before, all neural nets were simulated and therefore took a lot of computing power just to do small things like machine vision (face recognition, etc...).

A Bird IS a better machine..

[Anonymous Coward]

I have to say, Norvig's comment about building airplanes instead of birds because birds is not the goal is kind of ridiculous. Of course, if the goal is -just- to fly, then airplanes is one way. And if you want to fly people, you need different materials (metal), etc.

But birds are -much- more complex in so many ways, including flight itself. They use their wings both for lift (gliding) and for powered flight. Most can dynamically change the angle of attack, and can rotate each wing separately. Large wingspan birds are incredibly efficient. And this is not to mention their ability to see (vision), navigate, communicate (with other birds), etc.

Auto-pilot is now quite sophisticated, but airplanes don't form flocks (a nod to Univ.of.Penns quadrotors here), they don't roost, and they don't produce their own offspring. If you want to move people around quickly, the airplane is the -minimum- technology. But it certainly isn't close to intelligent in the ways that a bird is.

I'm working on it!

[Ramin_HAL9001]

Just give me a few more months, we'll have a HAL9000 soon enough.

Super computers may already beat humans

[jrincayc]

Depending on who you talk to, we already may have computers with enough processing power.

http://www.scientificamerican.com/article.cfm?id=computers-vs-brains
says that there are super computers with more processing power than humans (tho on a computations per watt humans are still much more efficient).

The Worldâ(TM)s Technological Capacity to Store, Communicate, and Compute Information by Martin Hilbert and Priscila LÃpez in (
Science 332, 60 (2011); DOI: 10.1126/science.1200970) states that the world had more computational power than a single human brain in 2007, but they basically assumed that every human neuron was operating at once, so their estimate was probably very conservative (unless human brains can somehow do non-neuronal processing).

humans neuter pets

[jrincayc]

I would like to make the comment, that even humans who claim to love their pets and consider them "people", think nothing of spaying and neutering them.

Five possible futures

[jrincayc]

1. Indistinguishable from magic.
2. No humans any more.
3. The sky is the limit (because the robot won’t let humans leave earth).
4. Thou shalt not create transistors.
5. Philosophical materialism is false.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

A simulation of a forest fire is not a forest fire.

Re:It's not just specialization, there is also fea

[Anonymous Coward]

(Clarification: The possibility of a difference between a thing and the simulation of that thing doesn't, in and of itself, prove that a simulation of a brain wouldn't produce a mind. However, assume that it necessarily would is a mistake, because a simulation of something doesn't necessarily have all the properties of the thing itself.)

Re:It's not just specialization, there is also fea

[mhajicek]

Nice. Categorical denial without even reading the article. Way to have an intelligent debate.

Re:It's not just specialization, there is also fea

[mhajicek]

Among other problems, you run into definitional issues, like "What do you mean by 'understand'?"

Exactly. There are different degrees of understanding. You and I may be able to discuss etymology and other abstract facets of language, which is a step above simply understanding the words. The robots in the article have only the most basic level of understanding; knowing that a word indicates a location, direction, or approximate amount of time. If put to a binary test of understand vs not understand, I'd say they understand. But they don't understand their understanding as we do; that's a higher level function.

Re:It's not just specialization, there is also fea

[mcgrew]

Children are born, not built. Your children are a continuation of you.

Re:It's not just specialization, there is also fea

[mcgrew]

Your brain is chemical, just like every other organ. Yes, there are ions and electrical charges, but you have that with any other chemical reaction.

It's likely that some time in the future we'll have Blade Runner-like "replicants" that would be sentient, but not Turing Archetecture electronic computers. If you understand how computers work, down to the level of logic gate circutry, you know that the only intelligence involved is the circut designer's and the programmer's intelligence.

Re:It's not just specialization, there is also fea

[mcgrew]

Silicon is also a chemical.

True, but it isn't chemistry that makes a computer work, it's electricity. The first computers had no silicon, they used vaccuum tubes; the transistor hadn't yet been invented.

Your computer is simply billions of on-off switches. Your brain is billions of different chemical reactions.

Re:It's not just specialization, there is also fea

[mcgrew]

I don't fear PETA, but I'd be pretty upset if they managed to get eating meat outlawed. What I fear is having rights taken away from me.

Re:It's not just specialization, there is also fea

[jpate]

What a pointless thing to argue. In the 1980 paper, he does indeed spend a great deal of time defending the proposition -- he doesn't provide the formal argument everyone is familiar with until later (1990), where it is indeed taken as axiomatic. However, as early as 1983 Searle writes "2. Syntax is not sufficient for semantics. That proposition is a conceptual truth. It just articulates our distinction between the notion of what is purely formal and what has content." from Minds, Brains, and Science pp. 28-41

So, yes, I stand by my assertion that the illustration is indeed a waste of time to argue about -- all that matters to the argument is the proposition.

I picked the earlier formulation because it is more defensible. The 1990 formulation just baldly asserts that "syntax is insufficient for semantics." Simply calling it axiomatic is not enough; you have to show why it is axiomatic. I'm open to the idea that a syntactic system cannot create its own semantics, but, if all we mean by semantics is an association between a sign and what it signifies, there's no reason that association could not be defined by some other syntactic system.

We get semantics from the visual system because the visual system provides a mental state (the sign) that corresponds with or represents something in the real world (the signified).

To the computer, there is no "real world" there is no distinction between data pulled into memory from a video camera or a stack of Hollerith cards nor from data already in memory or data being gathered at the time it's accessed -- there is no distinction. The computer is just manipulating meaningless symbols (and even that's a stretch, as the computer can't make such a distinction!) Meaningless symbols in relation to one another are ... meaningless symbols, being manipulated meaninglessly.

Well, to the human brain there is also no such thing as a "real world." Haven't you seen the matrix? ;) we don't experience the real world directly; brains interpret signals from our various senses.

Finally, I'll just point out you still haven't explained why neurons can cause a capacity for Chinese (or if you don't like the language examples, calculus, or baseball, or music.

The alternative is to posit a non-physical explanation. Searle doesn't deny that brains cause minds -- he only argues that what brains do to cause minds can not be computation alone. I don't have the answer, and neither does anyone else. I suspect that the answer will not come from philosophy or neuroscience, but from physics as a necessary consequence of some undiscovered bit of reality.

But this question is at the heart of the Chinese Room argument. In the 1990 presentation you favor, Searle says "if I do not understand Chinese solely on the basis of running a computer program for understanding Chinese, then neither does any other digital computer solely on that basis." In the argument, the man in the Chinese room is a substrate for the computation, just part of the machine. We don't expect individual neurons to understand Chinese, or individual transistors, so we should also not expect the man in the room to understand Chinese. The argument, and this was my original point, is just misleading. It doesn't show anything.

Of course, this is not going to influence your position, because you believe that it is self-evident that syntax is insufficient for semantics, and take, as Searle came to, the Chinese Room Argument as an illustration rather than an argument. I don't see it as self-evident at all, since, again, semantics is an association between a sign and the signified, and those associations are one kind of thing a syntactic system can compute. And with this, I will conclude my participation in this discussion. Good day.

Re:It's not just specialization, there is also fea

[Kielistic]

It does if you give it the properties. A mind is either a physical thing which can be simulated, emulated or recreated or it is a magical thing. I doubt it is a magical thing.

Re:It's not just specialization, there is also fea

[narcc]

Of course, this is not going to influence your position, because you believe that it is self-evident that syntax is insufficient for semantics

Indeed, your belief that Searle inadequately defends the assertion is irrelevant to me. "Searle didn't do a very good job of defending the claim" would be a very silly reason to reject for me to reject the premise! I'd need a real reason as the claim to me seems obvious on its face.

I don't see it as self-evident at all, since, again, semantics is an association between a sign and the signified, and those associations are one kind of thing a syntactic system can compute.

That's because you're confusing syntax with semantics. Redefining semantics in terms of syntax doesn't get you anywhere as we're no longer talking about the same thing. Relationships between symbols are indeed syntactic -- and as the dictionary example I offered you illustrates, can not give rise to semantics.

In the argument, the man in the Chinese room is a substrate for the computation, just part of the machine. We don't expect individual neurons to understand Chinese, or individual transistors, so we should also not expect the man in the room to understand Chinese.

Had you read a little further, you'd know that this was addressed already by Searle in the 1980 paper . Imagine now that our victim in the room get's so good at his job that he's memorized the rule book and can carry out his shuffling without reference to the book. Here, the entire process happens "inside his head". Still, the mere application of the rules is still not sufficient for meaning even though the whole operation is taking place inside a living brain!

If you need a simpler example, consider if he internalized a Chinese version of Weizenbaum's ELIZA program (which passes the Turing test to some degree -- ask his secretary) Add additional complexity to the rules slowly and you'll see that at no point can we introduce something new from which semantics can arise -- it's just more of the same meaningless manipulation of symbols.

Replies like the robot reply try to sneak semantic content "through the back door" as it were, by trying to trick us into introducing phenomenal experience. In the robot case, we have a camera in which our prisoner can see the external world. It fails, of course, when we replace the camera with an analog to what our brain gets -- more meaningless symbols. Now the situation is worse, not only can we see that the man won't learn Chinese, he can't even tell which slot introduces questions and which introduces visual information! (It's equally clear in the original that the man had no way to determine even something as simple as that he was getting questions and producing answers. Semantics at the very highest level of description are still unavailable to him!)

The only way out appears to be to introduce a non-computational component to "bridge the gap". Hence, syntax alone is insufficient for semantics.

(It doesn't take much of a leap here to make the claim that phenomenal experience is the missing piece requisite for semantic content. From which we can conclude that phenomenal experience can't be computational.)

And with this, I will conclude my participation in this discussion.

Okay... I had forgotten all about it days ago!

Re:It's not just specialization, there is also fea

[ChetOS.net]

Plenty? Please cite for me (in a peer-reviewed journal) even one experiment which demonstrates that we evolved into sentient beings from non-sentient organic matter.

Re:It's not just specialization, there is also fea

[ChetOS.net]

Please cite for me (in a peer-reviewed journal) even one experiment which demonstrates that we evolved into sentient beings from non-sentient organic matter.

Re:It's not just specialization, there is also fea

[nurb432]

you know that the only intelligence involved is the circut designer's and the programmer's intelligence.

If you include self modifying code ( which already exists ) then that statement isn't 100% as things will evolve beyond what the programmer wrote.

Also, what if you simulate the entire brain and its inputs/outputs? The simulated electrical activity inside the brain could be just as sentient as we are... ( i know we cant yet, but someday... )

Re:It's not just specialization, there is also fea

[mcgrew]

If you include self modifying code ( which already exists ) then that statement isn't 100% as things will evolve beyond what the programmer wrote.

I've written self-modifying code, but the code still operates under the conditions and constraints on it. It isn't thinking, it's just following instructions.

As to a simulation, a simulation of an atomic explosion produces simulated radiation, not real radiation. A simulated brain would likewise produce simulated thought.

You can't make real butter without cream. You can simulate it with hydrogenated oils, but margarine isn't butter, even if it will fool some people.