IMHO, some of the objections are kinda silly, especially "Wouldn't they be programmed with...?" We know so little about how the mind works, we have almost no idea what the constraints and requirements of AI would be. Furthermore, you can't wave a magic wand and say "My robot will do this, this, and this." Roger Clarke wrote "...the sum of Asimov's stories disprove the contention that he began with: It is not possible to reliably constrain the behavior of robots by devising and applying a set of rules." - Ununnilium
: It seems like you have two distinct sub-classes of the Ridiculously Human Robot
. What you're objecting to is "Ridiculously Human Emergent Behavior" — robots developing unprogrammed complex behaviors that mimic human behavior in an unpredictable way, and I think your complaint is mostly reasonable on this point (Though there really are very complex and hard-to-explain reasons why a lot of what we see really *is* unreasonable, that run along the line of the same reasons that I personally am not liable to sprout wings any time soon). The other category, "Ridiculously Human Design Features", is based on the notion that human
robot designers would think it was a good idea to give the robot a penis. Consider how often shows have had to reveal that the designer was a little crazy just to justify why they'd build a robot that way.
BT The P
- Wulfenbach: Attractively humanlike robot with penis can be used for sex or even romance. What's ridiculous about it? The very popularity of the genre, not to mention people who want to create human-like AI for one reason or another, indicates there'd be a market. And if vibrators, 2D porn, and inanimate creepy-looking sex dolls can sell, why not full-fledged companions?
: I think that Futurama
's robot society is like The Simpsons
' flexible geography. It's a subtle joke that is often too subtle to be funny. The omnipresent elements of roboculture are more a parody of this trope than an example of it, plus serve to lampoon the segregation of present day society in the media.
Ununnilium: Both good points. Don't forget, however, that many robot creators in TV are Mad Scientist
s, and thus a little crazy anyway.
: Incidentally, what Roger Clarke wrote was exactly what Asimov was trying to prove. It is possible to constrain
an AI's behavior, but not to contain
it within predictable specifications.
I'm going to be mildly contrarian here and suggest that there may be good reasons why human-intelligence robots would have to
possess human-mimicing features like emotions. First, the Spock
alternative is not realistic either. Your hypothetical completely unemotional AI would not be motivated to do anything
. Second, it's unclear (speaking here as someone who's read a ton of actual AI research) that it is possible to program
an AI; it may be necessary to grow
one rather as a child grows, and if you don't start it out with all the basic human capacities (including all the emotions) you probably get an autistic AI at best. — Zwol
: A good point. However, may of the AIs
in these fictional universes are explicitly programmed, and often, shown next to Spock-like AIs
that function just fine. I should put that in the entry, actually.
: Perhaps one should make a distinction between a true Artificial Intelligence and an Expert System. An AI is self-aware, and according to a Turing test, indistinguishable from a human participant. An Expert System on the otherhand, is merely a computer that is programmed in such a way that it seems
human, but is only designed to do a limited number (Usually just one) thing. A Library computer, for example, might be programmed to be fully versed in literature, art, etc...but if you strayed outside of that, you'd end up with canned responses like "I'm sorry, please repeat or rephrase your query". The Enterprise' computer in TNG is an Expert System..not self-aware (Except for that episode when it was). Data is
I watched I, Robot last night, so I'll use that as an example. The older mass-produced robots were for the most part Expert Systems. Programmed to interact with humans in a way that most would find comforting. The NS-5s were more advanced, but still mostly just very advanced Expert Systems. Sonny and VIKI however, were true AIs
. Sonny more than VIKI. They knew they had a "self", they knew they couldn't be easily replaced, that their experiences were unique. This is something that even humans
have to learn, and aren't capable of before age three or so.
(A little later) Oh, also, the hologram projector of Dr. Lanning is the perfect example of an Expert System. "I'm sorry, my responses are limited. You have to ask the right questions."
: Hmmmmmm, makes sense. But how do we tropeify it?
: And for a more philosophical question... Where is the line between artificial intelligence and actual intelligence? At what point can we consider a robot to be alive? And yes, the distinction between personality and programming is
important to this trope. There's a difference between a robot simply programmed to protect life and one that's aware enough to show compassion and a protective urge. Or, for that matter, one programmed to kill and one that's sapient and simply bloodthirsty.
: "even Megaman himself was not truly capable of pure free will." Where does it actually say that in the games? I've only run into it as fanon
Roland: I believe, although I may be utterly wrong, that part of what made X so special was his complete free will and capacity to learn. By contrast, Megaman was programmed with a particular persona and could not evolve beyond it.
: Yeah, while it's never stated explicitly in the games, I think it's fairly well implied that this is the case. If Mega Man
-generation robots are capable of free will, what makes reploids (whose entire schtick is that they've got free will and emotions) so special?
I guess you leave it out if you want to stick with strict canon
: Yeah, the "Mega Man
is not capable of free will" thing is implied by the fact that the Mega Man X
storyline specifically has X's free will as the reason he has to spend 100 years in a capsule being run through tests. There's nothing in the mainline Mega Man
canon that shows it; it's just implied.
: Makes sense. Editing in.
: Not sure that the Rozen Maiden
example fits — the dolls weren't built to kill each other off; that's aside effect of their creator/father Rozen's true goal: to give form and shape to Alice, the perfect little girl who lives in his head.
Each doll has a shard of that potential in herself, and thus they have to collect the Rosae Mystica
from the others to realize that goal, but in their creation, they were to be as close to perfectly human as a dollmaker could manage to make them.