First a little background...

I'm clipping this to extend a discussion on Wildcat's clip about Volvo and their new "automatic steering" (which makes me nervous.)

Separate from my concerns about the reliability of Volvo's programing for such a system, a fascinating discussion of Artificial Intelligence emerged.

It happens to relate to a few other social, technological, and conitive issues being discussed around clipmarks.

First, I think we should revisit John Searle's Chinese Room Argument. It should also be supplemented by considering some hilarious [url=/clipmark/6275E446-43E2-4A8D...

First a little background...

I'm clipping this to extend a discussion on Wildcat's clip about Volvo and their new "automatic steering" (which makes me nervous.)

Separate from my concerns about the reliability of Volvo's programing for such a system, a fascinating discussion of Artificial Intelligence emerged.

It happens to relate to a few other social, technological, and conitive issues being discussed around clipmarks.

First, I think we should revisit John Searle's Chinese Room Argument. It should also be supplemented by considering some hilarious examples of Chinglish that michellezm clipped which illustrate Searle's point.

Second, it's worth checking out Abailart's awesome clip on the role creativity plays in the retention of knowledge. Creativity is something dangerously underestimated by mechanically-oriented thinkers like Goertzel.

Finally, there are my thoughts on another organic-digital interface, namely the spread of Clipcasts, and it's role in the social aspect of potential cloud computing.

I would be fascinated to hear what anyone has to say.

Actually, this is very interesting that you also see the connection between experience, identity and intelligence on the one hand and creativity and non-linear thinking on the other.

From your profile it sounds like you have some experience with Autism. I wonder what your perspective is on this in relation to communication and creativity.

From what I understand, Autism, or at least Asperger's syndrome is that it seriously impairs creativity and abstract or holistic thinking, but not analytical thinking.

Now this strikes me as highly analogous to the problem we are dealing with in creating AGI, which is excellent analytically, but can only mimic abstract or creative thought.

Most provoc...

Actually, this is very interesting that you also see the connection between experience, identity and intelligence on the one hand and creativity and non-linear thinking on the other.

From your profile it sounds like you have some experience with Autism. I wonder what your perspective is on this in relation to communication and creativity.

From what I understand, Autism, or at least Asperger's syndrome is that it seriously impairs creativity and abstract or holistic thinking, but not analytical thinking.

Now this strikes me as highly analogous to the problem we are dealing with in creating AGI, which is excellent analytically, but can only mimic abstract or creative thought.

Most provocatively, computers also have serious problems taking initiative or developing communication independent of external commands. If we can figure out how to make computers take initiative in communication, it might provide revolutionary insights in improving communication skills among non-communicative autistics.

Now this strikes me as highly analogous to the problem we are dealing with in creating AGI, which is excellent analytically, but can only mimic abstract or creative thought.

Most provocatively, computers also have serious problems taking initiative or developing communication independent of external commands. If we can figure out how to make computers take initiative in communication, it might provide revolutionary insights in improving communication skills among non-communicative autistics.

My thoughts exactly. I have three children who are high functioning autistic and one with a preliminary diagnosis of PDD. We are currently partisipating in a couple of stu...

Now this strikes me as highly analogous to the problem we are dealing with in creating AGI, which is excellent analytically, but can only mimic abstract or creative thought.

Most provocatively, computers also have serious problems taking initiative or developing communication independent of external commands. If we can figure out how to make computers take initiative in communication, it might provide revolutionary insights in improving communication skills among non-communicative autistics.

My thoughts exactly. I have three children who are high functioning autistic and one with a preliminary diagnosis of PDD. We are currently partisipating in a couple of studies at Boston University.
At the age of two my daughter could read at a five year level and count to 100. But humor( which is non-linear) is frankly lost on her.

Okay, well, glancing at my book shelf, I think there are a few books I could throw your way. The questions we're dealing with seem to be: the mind, communication, social interaction, language, and atypical neurological function (within which I classify autism).

In general, I don't like walking in other people's footsteps, and I don't think you should either. So instead of theories of autism, here's a reading list that gets to the core of the issues and can help you form your own opinions on how the autistic mind works.

I feel they would be equally relevant to developing AI, since it should be noted that AI scholars are almost universally computer scientists with relatively little knowledg...

Okay, well, glancing at my book shelf, I think there are a few books I could throw your way. The questions we're dealing with seem to be: the mind, communication, social interaction, language, and atypical neurological function (within which I classify autism).

In general, I don't like walking in other people's footsteps, and I don't think you should either. So instead of theories of autism, here's a reading list that gets to the core of the issues and can help you form your own opinions on how the autistic mind works.

I feel they would be equally relevant to developing AI, since it should be noted that AI scholars are almost universally computer scientists with relatively little knowledge of the human mind.

Basically, I suggest a three-pronged approach.

1. Reading on the subject in popular culture and layman's terms, to help you develop an overall mental schema of the issue and help you organize the rest of the information you find.

2. Reading on primary sources, namely understanding theories of the mind, language, creativity and socialization. This way we go straight to the heart of the issue, rather than having to rely on the interpretations provide by professional mouth pieces of conventional wisdom.

3. Reading on unrelated material for comparative sources, which may provide useful analogies that serve as indirect methods of accessing the more difficult problems. This is key to coming up with a new situation rather than simply repeating institutionalized mistakes until we accept them.

thats just like the sci fi film minority report.

thats like psychic computing or prophetic computer systems.

I think philip k dick (sci fi author) came up with that idea a while ago

Yes,
it's been addressed in several sci-fi books. In the first book of the
Hitchhiker's Guide series of Douglas Adams there was such a computer,
which was just about to answer the question of the meaning of life...
well, then the world was destroyed as far as I remember... I didn't
think of it as an AI, because it wasn't self improving or so. It just
took (lots of) information, made calculations, and gave answers
(right?).

But what I thought of as AI, is apparently called AGI. So the computer ab...

thats just like the sci fi film minority report.

thats like psychic computing or prophetic computer systems.

I think philip k dick (sci fi author) came up with that idea a while ago

Yes,
it's been addressed in several sci-fi books. In the first book of the
Hitchhiker's Guide series of Douglas Adams there was such a computer,
which was just about to answer the question of the meaning of life...
well, then the world was destroyed as far as I remember... I didn't
think of it as an AI, because it wasn't self improving or so. It just
took (lots of) information, made calculations, and gave answers
(right?).

But what I thought of as AI, is apparently called AGI. So the computer above would be a narrow AI then...

That's
the kind of system I meant; and I was just wandering, if some work in this direction
is going on. I think with todays technology it is rather sicence than
fiction. At least the first step towards it... (I wouldn't be
surprised, if it tunrs out, that Google is working on this )

Minority report was a little bit different though: There were beings with special mental abilities involved.

tidbit2:
Can you remember the name of the book of Philip K. Dick? In remember
one of his books (Time out of Joint), in which a man could recognize
patterns in a puzzle, which were actually predictions about where the
next nuke from the moon will fall (or something like that). But you
mean one including a predicting computer I understand?

A while ago, I was discussing this topic with a
friend. He saw humans as "biological computers" and couldn't see any reason, why
mechanical beings (computers/robots) should not have our thinking and
feeling capabilities in the future. Yes, he thought that they may even have/show
emotions. I could by no means imagine that.. "Showing emotions" is maybe not so
difficult, you can program it to produce tears in certain situations. But
"feeling"? I mean, he/she/it could pronounce the words "you broke my heart",
and mimic a 'sad face', but would/could it really feel something? I still
cannot imagine, how this should be possible.
He (being an atheist) and I (being a believer, s...

A while ago, I was discussing this topic with a
friend. He saw humans as "biological computers" and couldn't see any reason, why
mechanical beings (computers/robots) should not have our thinking and
feeling capabilities in the future. Yes, he thought that they may even have/show
emotions. I could by no means imagine that.. "Showing emotions" is maybe not so
difficult, you can program it to produce tears in certain situations. But
"feeling"? I mean, he/she/it could pronounce the words "you broke my heart",
and mimic a 'sad face', but would/could it really feel something? I still
cannot imagine, how this should be possible.
He (being an atheist) and I (being a believer, so
to say) have different viewpoints. I believe in the existence of a soul, and
I think it is the soul (coupled with the body), that does the "actual feeling".
The body and the hormones etc. just produce the "viewable effects". But it's the
soul that gets hurt, not the body. And I cannot imagine a machine with a soul. Which doesn't mean, I believe it is not possible.
(I cannot but remember Data (Star Trek). He sat before an electric kettle, waiting 1lt of water boil several times, to get an idea about how humans feel the time passing slower/faster depending on what they're doing. He said something like "My chronometer measures the same time, whatever I do while waiting". Riker: "Humans don't have a chronometer, try turning it off".)

I am getting off the track, I realize. It wasn't about AGI's having emotions. (I hope, the off-topic part contributes at least to the "brain storming" going on here). To recur to the subject:

could Kurzweil and Goertzel's Aritifical General Intelligence genuinely surpass the human brain? Or is ti simply mimicing it?

Since they speak (wrote) about brain-computer interfaces, genetics and AI in the same sentence, my first thought was that they mean a computer somehow coupled with a human brain/being. But I should read the article first...

Just want to briefly mention an idea from a Tad Williams book (Otherland) first: There was a Virtual Reality (VR), in which people could "plu...

I am getting off the track, I realize. It wasn't about AGI's having emotions. (I hope, the off-topic part contributes at least to the "brain storming" going on here). To recur to the subject:

could Kurzweil and Goertzel's Aritifical General Intelligence genuinely surpass the human brain? Or is ti simply mimicing it?

Since they speak (wrote) about brain-computer interfaces, genetics and AI in the same sentence, my first thought was that they mean a computer somehow coupled with a human brain/being. But I should read the article first...

Just want to briefly mention an idea from a Tad Williams book (Otherland) first: There was a Virtual Reality (VR), in which people could "plug themselves in" via an "integrated interface". Later, people entering the VR began to fall in a coma, and actually died when they died in the VR (as we saw in The Matrix later). There were AIs in the VR (virtual beings), that could feel, cry, be happy, get hurt etc. Williams had presented it in such a way, that I was for the first time (at least) thinking about such a thing being possible. But (spoiler for those, who consider reading the 4-book-series) eventually it was revealed, that a child being able to manipulate peoples brains, was somehow linked with the computer (being its main part actually) that was generating/maintaning this VR.

@Djiezes: thank you for the links! I will have some great reading in the following days (I've been given a new Terry Pratchett book today, which adds to the joy).

So, to here's books I'd suggest for both the AI and the Autism question...

An Anthropologist on Mars
by Oliver Sacks
If you haven't already read this book, you shouldn't miss it. Oliver Sacks, a creative, sensitive, and humane thinker approaches a number of "illnesses" through anecdotes of his patients, including a very interesting exploration of autism in the eponymous final chapter.

[url=http://www.amazon.com/Exuberance-Passion-Kay-Redfield-Jamison/dp/0375701486/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1197993237&sr=1-1]Exuberance: The Passion for Li...

So, to here's books I'd suggest for both the AI and the Autism question...

An Anthropologist on Mars
by Oliver Sacks
If you haven't already read this book, you shouldn't miss it. Oliver Sacks, a creative, sensitive, and humane thinker approaches a number of "illnesses" through anecdotes of his patients, including a very interesting exploration of autism in the eponymous final chapter.

Exuberance: The Passion for Life
by Kay Redfield Jamison
A great book for exploring the mysterious forces of creativity and enthusiasm. Jamison is a brilliant, mentally ill psychiatrist who writes poetic, witty and heartbreakingly lucid books that provide valuable insight into the relationship that often exists between "illness" and "greatness."

Philosophy of Mind: Classical and Contemporary Readings
by David Chalmers
This is a great overview of the field of Philosophy of the Mind. It's edited by David J. Chalmers, who is an interesting guy in his own right. It is also one of the easier ways to jump into what is a highly esoteric field.

The Language Instinct: How The Mind Creates Language
by Steven Pinker
Pinker is in the class of scholars best described as "Monty Python-quoting geniuses who need a hair cut." In this book he makes the fascinating connection between people like his MIT colleague Noam Chomsky, who believe that language is a product of our neurological architecture, and those who believe that language is acquired fr

...and those who believe that language is acquired from our environment through interaction.

A Universe of Consciousness: How Matter Becomes Imagination
by Gerald M. Edelman
Bringing these all together is this interesting book by Nobel-prize winning biochemist and neurologist Gerald M. Edelman. Fair warning, this book can get pretty dry and technical, so don't bite off too much at once.

[url=http://www.amazon.com/Freedom-Neurobiology-Reflections-Political-Philosophy/dp/0231137524/ref=pd_bbs_sr_3?ie=UTF8&s=...

...and those who believe that language is acquired from our environment through interaction.

A Universe of Consciousness: How Matter Becomes Imagination
by Gerald M. Edelman
Bringing these all together is this interesting book by Nobel-prize winning biochemist and neurologist Gerald M. Edelman. Fair warning, this book can get pretty dry and technical, so don't bite off too much at once.

Freedom and Neurobiology
Mind: A Brief Introduction
The Construction of Social Reality
by John R. Searle
These three books should provide a fairly accessible introduction to Searle, one of the foremost thinkers on all these questions on the human mind, and a prominent critic of Artificial Intelligence.

Play, Sing & Dance: An Introduction to Orff Schulwerk
by Doug Goodkin
As a comparative example, consider classical composer Carl Orff's groundbreaking work on music education in children. I have also heard of it being used to improve communication skills among the autistic, and it is very interesting because music falls half way between the poles of abstract and analytical thinking that we are exploring.

Thankyou so much.

I understand what you mean about following others footsteps. It's kind of hard to look at things from a different angle when someone's telling you what you should be seeing.

I think as they develop AI they may come up with more ways of treating autism.
Autistic children have a have time when the encounter something that the have not experienced before or have not learned by ROTE. Which is IMO is just like the programming for AI. If AI encounters something that the programmer didn't account for , there is going to be a problem.

Take our Volvo, for example. The Volvo has a wonderful communication system going with accompanying traffic, and staying the prescribed distance...

Thankyou so much.

I understand what you mean about following others footsteps. It's kind of hard to look at things from a different angle when someone's telling you what you should be seeing.

I think as they develop AI they may come up with more ways of treating autism.
Autistic children have a have time when the encounter something that the have not experienced before or have not learned by ROTE. Which is IMO is just like the programming for AI. If AI encounters something that the programmer didn't account for , there is going to be a problem.

Take our Volvo, for example. The Volvo has a wonderful communication system going with accompanying traffic, and staying the prescribed distance behind a truck . Suddenly debris comes flying off from the back off the truck.
How will our Volvo handle this?
Will it be able to detect the debris?
Will it consider the option of swerving into the breakdown lane to avoid the debris?
Will it understand that to be acceptable under certain circumstances?

Djeizes:
Thanks for the tip, I'm checking out Nanay.

Abailart:
Take your time. Looking forward to your thoughts. I think the creativity question is actually a bigger bottleneck to the development of AI than any software problem. The main reason we can't program a computer to think like a human is that, honestly, we still don't understand how or why humans think.

Splendidus:
Well, actually, your friend illustrates the main problem in this, namely that computer scientists and mathematicians aren't neurobiologists, and real neurobiologists generally aren't interested in AGI.

The main problem, as I see it, is two fold: First is one of architecture, and the second is one of will.

The first...

Djeizes:
Thanks for the tip, I'm checking out Nanay.

Abailart:
Take your time. Looking forward to your thoughts. I think the creativity question is actually a bigger bottleneck to the development of AI than any software problem. The main reason we can't program a computer to think like a human is that, honestly, we still don't understand how or why humans think.

Splendidus:
Well, actually, your friend illustrates the main problem in this, namely that computer scientists and mathematicians aren't neurobiologists, and real neurobiologists generally aren't interested in AGI.

The main problem, as I see it, is two fold: First is one of architecture, and the second is one of will.

The first question is: How does the brain work?

The brain is not simply a biological computer, except for in the most basic of all layman's terms. A computer is digital. This means it is a system of circuits that go on or off, the synchronicity / pattern of these circuits is translated into binary which is then interpreted by a program. (This is on the most basic level, of course.)

The brain, on the other hand, although it does have neurons which may or may not resemble circuits, doesn't work like a computer at all. Essentially, we still don't know how the brain works, but it does so in a way fundamentally different from digital computing.

The other question is: Why does the brain work?

Creativity, free will, motivation, these are all things that allow our brain to function without outside input. Our brain can think, but what makes it start thinking. We can communicate, but why do we decide to do so?

Even sidestepping the spiritual questions, we can't simply ignore free-will because it is inconvenient. It would be much simpler if our mind simply followed a mathematical model with digital circuitry, but wishing will not make it so.

ljsdesign:

This is the interesting thing. A purely analytical brain, be it autistic or digital, often has no motivation to interact with others. Perhaps because it doesn't see the world that way.

Let's consider the Volvo for a moment, but change the scenario slightly. Instead of giving the Volvo a narrow AI for emergency response, suppose we give the Volvo a complete AGI (Like Knight Rider, but not as sassy, and with a Scandinavian accent, and maybe a little bit motherly.)

Presumably it would have its own perceptive mechanisms (something the narrow AI lacks) so it would be able to "see" the debris, and it might even consider acting to protect itself, since a self-aware computer would prob...

ljsdesign:

This is the interesting thing. A purely analytical brain, be it autistic or digital, often has no motivation to interact with others. Perhaps because it doesn't see the world that way.

Let's consider the Volvo for a moment, but change the scenario slightly. Instead of giving the Volvo a narrow AI for emergency response, suppose we give the Volvo a complete AGI (Like Knight Rider, but not as sassy, and with a Scandinavian accent, and maybe a little bit motherly.)

Presumably it would have its own perceptive mechanisms (something the narrow AI lacks) so it would be able to "see" the debris, and it might even consider acting to protect itself, since a self-aware computer would probably have a self-preservation instinct.

But the question is, what would it do? Would it be willing to break the rules, and do something wrong, because in this situation it seemed right? (ie. driving onto the shoulder.) This is subjectivity, which makes humans able to make decisions faster, although less precisely. In computer models, there is no real way to create genuine subjectivity yet.

Also, would the car act to preserve others as well? If so, why? Because it was programmed to? If that's the case, we're back at narrow AI, because it's not really thinking for itself, but just following a complex set of rules laid down by its programmer.

On the other hand, what is it that engenders empathy in the human brain? Empathy, which makes us want to connect positively with like beings... this is very difficult to calculate. Although the benefits of interaction seem obvious to us, it is difficult to translate into purely analytical terms, such as math.

Perhaps this is at the root of a lot of non-communication in Autistic children. It's not that they lack the physical or mental acuity to talk, it's just that they see no reason to do it.

And would the Volvo care about all cars? Or only other Volvos with AGI?

Splendidus:

Finally, getting around to your point about bio interface, this strikes me as a point where Goertzel and Kurzweil are a little confused about what they are talking about. They are day dreaming about making people (presumably themselves) super-intelligent, and using computers to eliminate death and work. They think AGI will be the solution to this, but they over look several fundamental problems in this slap-dash plan.

First, if we are talking about a computer to augment the human brain, it is not in fact AGI, but merely a tool used by our own native intelligence. Putting a self aware computer in my brain with a distinct identity would probably just mean it would nag me and figh...

Splendidus:

Finally, getting around to your point about bio interface, this strikes me as a point where Goertzel and Kurzweil are a little confused about what they are talking about. They are day dreaming about making people (presumably themselves) super-intelligent, and using computers to eliminate death and work. They think AGI will be the solution to this, but they over look several fundamental problems in this slap-dash plan.

First, if we are talking about a computer to augment the human brain, it is not in fact AGI, but merely a tool used by our own native intelligence. Putting a self aware computer in my brain with a distinct identity would probably just mean it would nag me and fight with me over what to watch on TV. I want a non-intelligent computer, or better still, just an internet connection so I can surf the web inside my head.

Second, we cannot enslave AGI computers to do work for us, because if we create genuine AGI, then we will have to treat them as autonomous and equal beings with ourselves, who are probably no more interested in work than we are, perhaps even less so!

Third, and most importantly, if we are telling the computer what to do, then it is not AGI. Thus bio-interface is not a true route to AGI, it is just sneaking human intelligence in through the back door. What we need to develop is a system of non-digital circuits that are operated in a new, analog way, creating a computer that works like the human brain.