I remember Asimov's story far more emotionally; but then, I was a mere teenager when I first read Bicentennial Man.
The short story has no expressed romance between Little Miss and Andrew Martin, as there is in the movie, and I'm sure it was better that way, because that just heightened the pathos: surely the very core of this story. Instead the film is forced to invent a lookalike granddaughter Portia in order to compensate for the robot's long lifespan. Cheesy.
I was disappointed in this, I have to admit. The reactions of the wife/mother Ma'am(Wendy Crewson, a wonderful actress under normal circumstances) are especially irritating, and frankly, not believable. Women, of all people, would accept robots far more readily and respond to them emotionally. For exactly the same reason that Little Miss liked Andrew almost instantly, any mother could and would muster--at least--warmth for him, ESPECIALLY if he was as gentle with the children as Andrew clearly was.
Remember Sarah Connor, who decided that she would let The Terminator raise her one and only child because he would make the best father of all the candidates? Well, if she could think like that (despite being highly suspicious), then why couldn't Ma'Am? It's crazy, and not believable.
The always-under-foot syndrome, too, could've been handled with more subtlety or just minimalist comedy, to show intelligent adjustment. Did the mother want to perform household chores herself or not? I completely fail to understand the woman's emotional/intellectual wherewithal (ineptitude).
Basically, we're forced to conclude that except for Sir and Little Miss, the rest of the family were lost causes. This despite that robots were supposed to be commonplace.
This abysmal character development has to be inherent to the Nicholas Kazan screenplay, we can't blame the unfortunate Wendy Crewson. The very same thing is wrong with AI(2001). The mother, again(?!), is even more insipid and daft. Such characters do test one's patience, because they are such an obvious fraud, for the sake of some plot-development or other: it is CLUMSY WRITING.
The Martins actually had a helpful, and an astonishingly verbally adroit android (compared to current technology), amazingly well-socialized into the family. He was safe (for them), and in his own way funny, courteous, smart, talented, and permanently so! I'd trade him in for any of my last three boyfriends. Why wouldn't any of these people come to terms with Andrew? Ridiculous.
Perhaps the problem was that Chris Columbus et al waited too long to make the movie. Other sci-fis like Terminator(1984), in fact, had already surpassed Asimov's notions about humanity's political incapacity and rejection of androids/robots. We've also (suffered through) I-don't-know-how-many "court cases" about android rights on Star Trek, and the Voyager Doctor's frankly irritating harping about hologram rights ad nauseam, so that declaring a robot as in fact human is no longer such a biggie. That had been the guts of the original 1976 story, but audiences are now far too sophisticated - watched 14yrs of Star Trek!
(Viewers interested in more Asimov robot stories should read "I, Robot" and "The Rest of the Robots". They are excellent short stories.) What I think happened was that certain contractual obligations about movie-making, and particularly this script, may have gotten in the way of creative output (eg the option was about to run out, and since the CGI technology was OK, it was now or never). We may have ended up with an out-evolved turkey because of the business being "not show-art". I don't know that this is what happened, but it sounds plausible to me.
The Three (really Four, Asimov added another one) Laws of Robotics is a very good place to start robot ethics; they are not in debate. The hard part about making Andrew real would be socializing him to the degree he was in the movie. That's what we still can't do. We can't teach context. That's the stumbling block. Neural nets do their learning usually with human cheating (called back-propagation - don't ask, I always thought the technique completely disreputable, so I cringed and winced may way through 7 years of Data's pride in his own neural net on ST-NG). But context requires generalization into abstracts from the concrete, and then reversing back again, and the mapping onto the real world has to be accurate >95% of the time. So it's a whole new kettle of fish. Roger Schank modularized layers of contexts as individual nested subroutines, and called them schemas, which a computer could simulate. But you just couldn't schematize/digitize enough of reality, it takes too much room. The human brain excels at filling in blanks (based on this bi-directional mapping of the "essential" & the "generalized"). Computers and robots suck at this. They have no real life experience to guess with. They need a background library the size of a planet to carry around just to take a walk in the park. Which is why Andrew Martin is so impressive. So. Wherefore art thou, Bicentennial Man? Perhaps Rodney Brooks (bottom-up architecture with Cog, Kismet) needs to team up with Doug Lenat (CYClist, top-down architecture, hand-codes reality one step at a time) so they can hit paydirt somewhere in the middle. That they should join forces is EXTREMELY unlikely: they are both using diametrically opposed concepts and code. The other possibility is that Honda will beat them to it, with their human-shaped white robot that walks up stairs. Otherwise, Bicentennial Man may be another couple of centuries away.
Growing up is aardvark (hard work).
The short story has no expressed romance between Little Miss and Andrew Martin, as there is in the movie, and I'm sure it was better that way, because that just heightened the pathos: surely the very core of this story. Instead the film is forced to invent a lookalike granddaughter Portia in order to compensate for the robot's long lifespan. Cheesy.
I was disappointed in this, I have to admit. The reactions of the wife/mother Ma'am(Wendy Crewson, a wonderful actress under normal circumstances) are especially irritating, and frankly, not believable. Women, of all people, would accept robots far more readily and respond to them emotionally. For exactly the same reason that Little Miss liked Andrew almost instantly, any mother could and would muster--at least--warmth for him, ESPECIALLY if he was as gentle with the children as Andrew clearly was.
Remember Sarah Connor, who decided that she would let The Terminator raise her one and only child because he would make the best father of all the candidates? Well, if she could think like that (despite being highly suspicious), then why couldn't Ma'Am? It's crazy, and not believable.
The always-under-foot syndrome, too, could've been handled with more subtlety or just minimalist comedy, to show intelligent adjustment. Did the mother want to perform household chores herself or not? I completely fail to understand the woman's emotional/intellectual wherewithal (ineptitude).
Basically, we're forced to conclude that except for Sir and Little Miss, the rest of the family were lost causes. This despite that robots were supposed to be commonplace.
This abysmal character development has to be inherent to the Nicholas Kazan screenplay, we can't blame the unfortunate Wendy Crewson. The very same thing is wrong with AI(2001). The mother, again(?!), is even more insipid and daft. Such characters do test one's patience, because they are such an obvious fraud, for the sake of some plot-development or other: it is CLUMSY WRITING.
The Martins actually had a helpful, and an astonishingly verbally adroit android (compared to current technology), amazingly well-socialized into the family. He was safe (for them), and in his own way funny, courteous, smart, talented, and permanently so! I'd trade him in for any of my last three boyfriends. Why wouldn't any of these people come to terms with Andrew? Ridiculous.
Perhaps the problem was that Chris Columbus et al waited too long to make the movie. Other sci-fis like Terminator(1984), in fact, had already surpassed Asimov's notions about humanity's political incapacity and rejection of androids/robots. We've also (suffered through) I-don't-know-how-many "court cases" about android rights on Star Trek, and the Voyager Doctor's frankly irritating harping about hologram rights ad nauseam, so that declaring a robot as in fact human is no longer such a biggie. That had been the guts of the original 1976 story, but audiences are now far too sophisticated - watched 14yrs of Star Trek!
(Viewers interested in more Asimov robot stories should read "I, Robot" and "The Rest of the Robots". They are excellent short stories.) What I think happened was that certain contractual obligations about movie-making, and particularly this script, may have gotten in the way of creative output (eg the option was about to run out, and since the CGI technology was OK, it was now or never). We may have ended up with an out-evolved turkey because of the business being "not show-art". I don't know that this is what happened, but it sounds plausible to me.
The Three (really Four, Asimov added another one) Laws of Robotics is a very good place to start robot ethics; they are not in debate. The hard part about making Andrew real would be socializing him to the degree he was in the movie. That's what we still can't do. We can't teach context. That's the stumbling block. Neural nets do their learning usually with human cheating (called back-propagation - don't ask, I always thought the technique completely disreputable, so I cringed and winced may way through 7 years of Data's pride in his own neural net on ST-NG). But context requires generalization into abstracts from the concrete, and then reversing back again, and the mapping onto the real world has to be accurate >95% of the time. So it's a whole new kettle of fish. Roger Schank modularized layers of contexts as individual nested subroutines, and called them schemas, which a computer could simulate. But you just couldn't schematize/digitize enough of reality, it takes too much room. The human brain excels at filling in blanks (based on this bi-directional mapping of the "essential" & the "generalized"). Computers and robots suck at this. They have no real life experience to guess with. They need a background library the size of a planet to carry around just to take a walk in the park. Which is why Andrew Martin is so impressive. So. Wherefore art thou, Bicentennial Man? Perhaps Rodney Brooks (bottom-up architecture with Cog, Kismet) needs to team up with Doug Lenat (CYClist, top-down architecture, hand-codes reality one step at a time) so they can hit paydirt somewhere in the middle. That they should join forces is EXTREMELY unlikely: they are both using diametrically opposed concepts and code. The other possibility is that Honda will beat them to it, with their human-shaped white robot that walks up stairs. Otherwise, Bicentennial Man may be another couple of centuries away.
Growing up is aardvark (hard work).