Slashdot Mirror
Digital People: From Bionic Humans to Androids
Aeonite writes "Cybernetics (that is to say, the sort associated with Cyberpunk) has long been an interest of mine, and so I was eager to dive into Digital People: From Bionic Humans to Androids , which is about exactly what it proclaims to be about. Sort of. The book does indeed cover everything from Bionic Humans to Androids, but the continuum of artificial beings is heavily weighted towards one end of the spectrum. Overall, the book is quite comprehensive in dealing with physical aspects of artificial intelligences, but when it comes to the nature of intelligence itself, the book barely dips its toe in the water. As the author himself says, the question here is not "Can machines think?", like Turing, but rather 'Can machines live?'" Read on for the rest of Aeonite's review.
Digital People: From Bionic Humans to Androids
author
Sidney Perkowitz
pages
248
publisher
Joseph Henry Press
rating
7
reviewer
Michael Fiegel
ISBN
0309089875
summary
A detailed exploration of the history of artificial beings
Where Have We Been? The book opens with a 13-page introduction ("Androids All Around Us") that sets up the material to follow, exploring the nature of artificial things and how we bond to them, personifying everything from the obviously inanimate cars and computers, to seemingly lifelike robots like MIT's Cog and Kismet, Sony's AIBO and Tiger's I-Cybie. The lines between what's alive and what's artificial are blurred even further when one considers that some 10 percent of the U.S. population are by definition bionic, possessing some degree of artificial parts -- everything from prosthetic limbs, to artificial hearts and hips, to breast implants and hearing aids. Not that this is anything new; the author tells us, for instance, that Aristotle imagined artificial beings in the 4th Century BCE, and several of the Norse gods and heroes had artificial hair and limbs. The difference now is that we're on the verge -- how close is up for debate -- of creating artificial, intelligent life, and it is the author's argument that now is the time to start thinking more about what that means. "To create artificial minds and bodies," he says, "we must first better understand ourselves."
The bulk of the book is split into two parts. The first, "Artificial Beings: Meaning and History" is semi-self explanatory, featuring three chapters that cover the history of robots, automatons and the like, from ancient times to the present day.
Chapter 2, "The Virtual History of Artificial Beings", is devoted entirely to fictional beings -- not only R2D2, Robocop, The Six Million Dollar Man, and other modern examples, but also Mary Shelley's Frankenstein, Pygmalion's ivory statue, the bronze Talos, Jewish golems and Frank L. Baum's Tin Man and Tik-Tok, among others. Interesting tidbits abound here: for example, the revelation that the term Robot -- first featured in Karel Capek's R.U.R. (Rossum's Universal Robots) -- comes from the Czech word "robota," which means "forced labor." Also interesting is the acknowledgment that for all the attention paid to Asimov's three laws of robotics, many of his stories deal with situations where those laws are broken, bent or otherwise shown to be invalid.
Chapters 3 and 4, on the other hand, cover real examples of artificial beings and bionic devices, from classical times to the early 1990s. Everything from Jacques de Vaucanson's musicians and Pierre Jaquet-Droz's automata, to explorations with galvanism in the 18th Century, to the development of computers and robots through the first half of the 19th Century are addressed. Again, the author digs up some fascinating insights and revelations here: mention of iron prosthetic legs in ancient Indian poetry; wooden and bronze legs for Greek and Roman soldiers; and the idea that advances in medicine since World War 2 have allowed soldiers to live with grievous wounds more often, which in turn has led to an increased need for development of better prosthetics.
Where Are We Now? The second part of the book, "How Far Along Are We?", spans five chapters, and covers mind-body interfaces, methods of robotic locomotion, sensory input, self-awareness and the like.
Chapter 5, "Mind-Body Problems", is very theoretical, offering more questions than answers, particularly as compared to the respectively crunchy material already covered. Can an artificial brain support a conscious artificial mind? Is a soul something special and unique to humans, or just the collective perceptions of a 3-pound mass of watery tissue? The following two chapters are a bit more crunchy, covering "Limbs, Movement and Expression" and "The Five Senses, and Beyond", respectively. Both provide plenty of concrete examples of the subject matter, with today's most advanced robots -- Honda's P2, P3 and ASIMO; Tokyo IT's snakelike ACM R-1; NASA's Robonaut, Spirit and Opportunity; and MIT's Kismet -- covered in reasonable depth. Problems and limitations of robots are also dealt with, including sensory bandwidth limitation, facial detection failure rates, the concept of "good enough" speech recognition, etc.
Chapter 8 takes us upstairs, covering "Thinking, Emotion and Self-Awareness" and the basics of what it means to have a robot brain. The world's three "smart" digital beings are covered in some depth here: ASIMO, the oldest, can walk backwards, keep balance, react to body language, recognize its name and wave hello; Kismet, who consists of only a head and face, can react to movements and expressions, but requires 15 networked computers to do so; and Sony's QRIO is only 23 inches tall, but can have 20,000 word conversations, sing in harmony, and greet people it knows by name, based only on facial recognition. Also covered are Commander Data's emotion chip, monkeys controlling robot arms with their brains, and the infamous light-seeking eels, ever a favorite on Slashdot.
Chapter 9, "Frankenstein's Creature or Commander Data", explores the ramifications of robotic development, and how they differ from society to society. In Japan, where roughly half of the world's million robots reside, artificial creatures have a mostly civilian role, whereas in America, military applications have a larger role. Does this result from religious differences? The author, citing Robert Christopher, suggests that Buddhists take a different view of robots than do Christians because Buddhism "does not place man at the center of the universe, and in fact, makes no particular distinction between the animate and the inanimate." Samurai swords have souls, and machines have ghosts.
Where Does That Leave Us? What, in the end, does this mean for us? Can machines be truly human if they never grow up, have no accumulated cultural experience? What does it mean for humans when the cost of labor rises while the cost of robots falls? Will it happen in our lifetimes? Turing predicted that a machine would pass his infamous test by the end of the century; Kurzweil says it will happen by 2029. Who's to say? Not the author -- he leaves off with no conclusions but that the journey will be uplifting, and will give us a sense of wonder at what we might accomplish.
This seeming lack of conclusion leaves the book a bit shallow, though one can't truly fault the author for not answering such a difficult question, especially since he backs away from "going there" at several points in the book. The author's refusal to speculate deeply about such matters make it clear why certain examples were "missing" from Part 1. While relatively minor works such as Marge Piercy's 1991 "He, She and It" were covered there, William Gibson's Neuromancer was notably absent (the author is only mentioned once, in passing, on page 189). And in a section that covered Blade Runner, The Terminator, Robocop and the Six Million Dollar Man, where were Ghost in the Shell and Max Headroom, both of which cover the nature of what it means to be artificial? The answer can probably be found in chapter 5's final sentence, which reads (in part): "...although the full mind-body recipe remains unknown for us and our artificial kin, a great deal of progress has been made on the bodily ingredient..." In other words, "we don't know much about the mental stuff, so let's look at the physical." This particular focus means that the book skews heavily towards a discussion of robots and robotics, with comparatively little attention paid to bionics and cybernetics; a better subtitle might have been "From Robots to Androids".
Also a bit troublesome is the fact that several areas -- particularly those dealing with more recent developments -- are glossed over, mentioned briefly, even tantalizingly, and then left behind. Electro-Active Polymers and the AMRI (Artificial Muscle Research Institute) are mentioned only in passing, and although brain-machine interfaces are mentioned several times, it's never with any real depth. The book's Filmography suffers from this focus on the past as well; only two of the 23 films and TV shows listed are from the past decade, with Star Trek and The Terminator left to represent the 1990s all alone. Although the book's historical perspective is intriguing, I would have preferred to hear a bit more about current events and examples.
Overall, Digital People is an enjoyable read, and is heavy with substance for those interested in learning about the history of artificial beings and robots, from ancient times to the 1990s. Those looking for more about cybernetics and human-machine interfaces might find themselves wanting more, but if your own tastes run more towards Asimov than Gibson, you won't be disappointed.
You can purchase Digital People: From Bionic Humans to Androids from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Where Have We Been? The book opens with a 13-page introduction ("Androids All Around Us") that sets up the material to follow, exploring the nature of artificial things and how we bond to them, personifying everything from the obviously inanimate cars and computers, to seemingly lifelike robots like MIT's Cog and Kismet, Sony's AIBO and Tiger's I-Cybie. The lines between what's alive and what's artificial are blurred even further when one considers that some 10 percent of the U.S. population are by definition bionic, possessing some degree of artificial parts -- everything from prosthetic limbs, to artificial hearts and hips, to breast implants and hearing aids. Not that this is anything new; the author tells us, for instance, that Aristotle imagined artificial beings in the 4th Century BCE, and several of the Norse gods and heroes had artificial hair and limbs. The difference now is that we're on the verge -- how close is up for debate -- of creating artificial, intelligent life, and it is the author's argument that now is the time to start thinking more about what that means. "To create artificial minds and bodies," he says, "we must first better understand ourselves."
The bulk of the book is split into two parts. The first, "Artificial Beings: Meaning and History" is semi-self explanatory, featuring three chapters that cover the history of robots, automatons and the like, from ancient times to the present day.
Chapter 2, "The Virtual History of Artificial Beings", is devoted entirely to fictional beings -- not only R2D2, Robocop, The Six Million Dollar Man, and other modern examples, but also Mary Shelley's Frankenstein, Pygmalion's ivory statue, the bronze Talos, Jewish golems and Frank L. Baum's Tin Man and Tik-Tok, among others. Interesting tidbits abound here: for example, the revelation that the term Robot -- first featured in Karel Capek's R.U.R. (Rossum's Universal Robots) -- comes from the Czech word "robota," which means "forced labor." Also interesting is the acknowledgment that for all the attention paid to Asimov's three laws of robotics, many of his stories deal with situations where those laws are broken, bent or otherwise shown to be invalid.
Chapters 3 and 4, on the other hand, cover real examples of artificial beings and bionic devices, from classical times to the early 1990s. Everything from Jacques de Vaucanson's musicians and Pierre Jaquet-Droz's automata, to explorations with galvanism in the 18th Century, to the development of computers and robots through the first half of the 19th Century are addressed. Again, the author digs up some fascinating insights and revelations here: mention of iron prosthetic legs in ancient Indian poetry; wooden and bronze legs for Greek and Roman soldiers; and the idea that advances in medicine since World War 2 have allowed soldiers to live with grievous wounds more often, which in turn has led to an increased need for development of better prosthetics.
Where Are We Now? The second part of the book, "How Far Along Are We?", spans five chapters, and covers mind-body interfaces, methods of robotic locomotion, sensory input, self-awareness and the like.
Chapter 5, "Mind-Body Problems", is very theoretical, offering more questions than answers, particularly as compared to the respectively crunchy material already covered. Can an artificial brain support a conscious artificial mind? Is a soul something special and unique to humans, or just the collective perceptions of a 3-pound mass of watery tissue? The following two chapters are a bit more crunchy, covering "Limbs, Movement and Expression" and "The Five Senses, and Beyond", respectively. Both provide plenty of concrete examples of the subject matter, with today's most advanced robots -- Honda's P2, P3 and ASIMO; Tokyo IT's snakelike ACM R-1; NASA's Robonaut, Spirit and Opportunity; and MIT's Kismet -- covered in reasonable depth. Problems and limitations of robots are also dealt with, including sensory bandwidth limitation, facial detection failure rates, the concept of "good enough" speech recognition, etc.
Chapter 8 takes us upstairs, covering "Thinking, Emotion and Self-Awareness" and the basics of what it means to have a robot brain. The world's three "smart" digital beings are covered in some depth here: ASIMO, the oldest, can walk backwards, keep balance, react to body language, recognize its name and wave hello; Kismet, who consists of only a head and face, can react to movements and expressions, but requires 15 networked computers to do so; and Sony's QRIO is only 23 inches tall, but can have 20,000 word conversations, sing in harmony, and greet people it knows by name, based only on facial recognition. Also covered are Commander Data's emotion chip, monkeys controlling robot arms with their brains, and the infamous light-seeking eels, ever a favorite on Slashdot.
Chapter 9, "Frankenstein's Creature or Commander Data", explores the ramifications of robotic development, and how they differ from society to society. In Japan, where roughly half of the world's million robots reside, artificial creatures have a mostly civilian role, whereas in America, military applications have a larger role. Does this result from religious differences? The author, citing Robert Christopher, suggests that Buddhists take a different view of robots than do Christians because Buddhism "does not place man at the center of the universe, and in fact, makes no particular distinction between the animate and the inanimate." Samurai swords have souls, and machines have ghosts.
Where Does That Leave Us? What, in the end, does this mean for us? Can machines be truly human if they never grow up, have no accumulated cultural experience? What does it mean for humans when the cost of labor rises while the cost of robots falls? Will it happen in our lifetimes? Turing predicted that a machine would pass his infamous test by the end of the century; Kurzweil says it will happen by 2029. Who's to say? Not the author -- he leaves off with no conclusions but that the journey will be uplifting, and will give us a sense of wonder at what we might accomplish.
This seeming lack of conclusion leaves the book a bit shallow, though one can't truly fault the author for not answering such a difficult question, especially since he backs away from "going there" at several points in the book. The author's refusal to speculate deeply about such matters make it clear why certain examples were "missing" from Part 1. While relatively minor works such as Marge Piercy's 1991 "He, She and It" were covered there, William Gibson's Neuromancer was notably absent (the author is only mentioned once, in passing, on page 189). And in a section that covered Blade Runner, The Terminator, Robocop and the Six Million Dollar Man, where were Ghost in the Shell and Max Headroom, both of which cover the nature of what it means to be artificial? The answer can probably be found in chapter 5's final sentence, which reads (in part): "...although the full mind-body recipe remains unknown for us and our artificial kin, a great deal of progress has been made on the bodily ingredient..." In other words, "we don't know much about the mental stuff, so let's look at the physical." This particular focus means that the book skews heavily towards a discussion of robots and robotics, with comparatively little attention paid to bionics and cybernetics; a better subtitle might have been "From Robots to Androids".
Also a bit troublesome is the fact that several areas -- particularly those dealing with more recent developments -- are glossed over, mentioned briefly, even tantalizingly, and then left behind. Electro-Active Polymers and the AMRI (Artificial Muscle Research Institute) are mentioned only in passing, and although brain-machine interfaces are mentioned several times, it's never with any real depth. The book's Filmography suffers from this focus on the past as well; only two of the 23 films and TV shows listed are from the past decade, with Star Trek and The Terminator left to represent the 1990s all alone. Although the book's historical perspective is intriguing, I would have preferred to hear a bit more about current events and examples.
Overall, Digital People is an enjoyable read, and is heavy with substance for those interested in learning about the history of artificial beings and robots, from ancient times to the 1990s. Those looking for more about cybernetics and human-machine interfaces might find themselves wanting more, but if your own tastes run more towards Asimov than Gibson, you won't be disappointed.
You can purchase Digital People: From Bionic Humans to Androids from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Do they dream of electric sheep?
Starsucks
Seems to me that the book will most likely pose more questions than it answers. Although, the type of questions it may pose may likely be the book's most entertaining aspect.
This is an implicit point of Dijkstra's quote:
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim."
Am I open minded towards open source, or closed minded towards closed source?
ouch, pun not intended.
-- I'm not a pessimist, I'm a realist. It's not my fault that life sucks so much. --
I think it's more important to know your cyborg name for when we welcome our new overlords.
A.N.D.R.E.W.: Artificial Networked Destruction and Rational Exploration Worker
Awesome.
Just refresh the page.
What they ended up with was something that sort of fakes human interaction. That's been done before. Remember Ananova? Chatterbots? My Real Baby, from Hasbro? COG is basically similar, but with a bigger budget.
The COG web site apparently hasn't been updated since 2000. Like the Leg Lab, it seems to have reached the limits of the ideas used.
This is sad, because there were some good ideas there. But they weren't anywhere near enough to even consider going to human-level AI in one jump. This is a classic vice of AI researchers - they have a reasonably good idea, and then start claiming that human-level strong AI is right around the corner. We went though this with the "expert systems" crowd in the 1980s, and that was even more embarassing and expensive, because doomed startups were launched. AI as a field was dead for a decade after that.
That's the price of overhyping a technology.
Well, I guess we had to go somehow. Its only a matter of time before some terrorist programmed super-bot decides to off us all. 1970 - Cold War 2001 - Terrorist Attacks 2021 - Reverse engineered Nanny-Droids I'll do my part now and wish our kind well, we did have a pretty good run. I wonder if Slashdot will be considered historical data to the new robot civilization? *Please access file 00342: Slashdot: A comparitive study of meat sacks, and their opinions. Later.
"Life" is a slippery, self-defined term. We know what our life is, as humans, more or less. But the "less" part makes it impossible to translate to forms other than the ones with which we're familiar. Are viruses (RNA) alive? Are prions, which are selfreplicating proteins? How about mules, which don't replicate: they're alive. "Life" is a term for chemical processes with some degree of complexity, some degree of spatial segregation, some degree of energy processing. Those degrees vary too much for a rigorous definition, because life transcends limits.
The definition of intelligence is more interesting, and still more relevant. We don't need other intelligences to be necessarily "smart" for them to be interesting to us. Intelligence is merely any degree of representation of the world, both inside and outside the thing, within the thing, which can be changed by changes in the world. In other words, anything that has a model of the world, including itself in the world, that dynamically changes corresponding to changes in the world, is intelligent. So computers are already intelligent. Just like fish are already intelligent. Some people who believe destroying a fish's intelligence by killing it might also believe that turning off a computer is killing the computer. They're probably right.
These debates show that we've now filled the natural world so much with human invention that we're bumping into our own creations, and trying to sort them out according to rules we made for creations we found before we could make our own. As long as we don't update our models to account for the new orders in the world, we're losing our own claims to "intelligent".
--
make install -not war
"MIT betrayed all of its basic principles."
It seems insane for anyone to go about attempting a "human-level strong AI" when there are much more basic levels that could be mastered. I would love to see some good ant AI first. I don't mean " capture the flag" ant AI... i mean build-an-ant-hill, fight-the-other-ant-colony, protect-the-queen ant AI. I bet there's some decent software versions of said AI, but i'd like to see it done with ant robots (i.e. - mini robot with feelers, mandibles, etc). Now that would be a doozy - managing communication via feeler tapping instead of by phat pipe. I would pay to see that.
(seriously.)
Also interesting is the acknowledgment that for all the attention paid to Asimov's three laws of robotics, many of his stories deal with situations where those laws are broken, bent or otherwise shown to be invalid.
First of all, Asimov's "three laws of robotics" aren't "laws" in any stretch of the imagination. They were only a plot device in a series of really good books. However, if we do start building robots to which the laws could apply, we would do well to keep them in mind. They should instead be called Asimov's "three very good suggestions to avoid building homicidal robots."
Second of all, the laws would only apply to robots with positronic brains, which would give said robots consciousness and the ability to make decisions. They would not need to be (and most likely could not be) applied to robots controlled by other humans via remote control, programming, or some other input (e.g. - every robot currently in existence). So the laws could not be properly applied to any robot in existance today.
I think this is a case of the writer and/or reviewer taking a concept well-known to the target audience, and then trying to get attention by creating a false sense of controversy about it. However, when taken in context (the three "laws" are a small part of a fictional world), it doesn't really matter. Sorry, but no laws broken here.