OK, what is it about human-like robots that makes people happy, in media and as a hypothetical reality?
1) Malfunctions? Subtle malfunctions? Extreme malfunctions? But what do malfunction do?
2) Body Panels and articulations? Again, why are these appealing?
3) Impersonation. A robot seeming to be a particular person or simply seeming to be fully human with a past, a role, etc? What is appealing about this
4) Robot movement? What sort and what's the attraction?
5) Control and programmability? What level of control?
6) Uncanny Valley? Being almost human but not quite. What makes that appealing?
7) Perfection? Being a built thing she may be able to perfectly satisfy some criteria
8) Free willed android or human mind in an android body? What's the appeal?
Personally, I prefer non-free-willed androids that impersonates a person but with a slight uncanny value quality. Malfunctions and body panels appeal to me only in as much as they reinforce the uncanny valley effect; example Carol Van Sant malfunctioning at the party in The Stepford Wives illustrates this.
But what appeals about the uncanny/impersonation thing to me? It's odd combination of fear and power I think. Beyond that, further mysteries.
So Which Robot Qualities Are Most Appealing?
Re: So Which Robot Qualities Are Most Appealing?
Interesting question, and one that makes you think about what it is about this topic that you really like.
1) I prefer subtle malfunctions, under the surface of the robot. Software and hardware control specific parts of the robot, so one of those malfunctioning would impact how that part works. The real interest is the tension that it creates when a failing system impacts what the robot is trying to do.
2) Body panels and articulations are fine. Panels should have a purpose, such as accessing specific parts of the robot. A robot that's indistinguishable from human may as well just be human.
3) I prefer robots that know what they are, even if some of them want to blend in among humans. I'd rather that they have their own past rather than borrowing one from someone else.
4) Robot movement isn't bad, but it should have a purpose. For example, stiff movement could indicate trying to conserve a low battery, or slower software emulation replacing a failed hardware motor controller. It would be an outward indicator of something specific going wrong underneath, or maybe just of a cheaply built robot.
5) Control is interesting because it's related to how a robot controls itself. The AI generates tasks that it wants to perform and uses programs and hardware drivers on itself as tools. The robot's state could directly impact tasks, like a task to connect an active charger getting higher priority as their battery gets lower. A command remote could create new tasks with a higher priority, or an intrusion would create tasks that appear to be genuine. Both would be indistinguishable from tasks the robot generated themself. Similarly, an intruder could insert or replace "tools" without the robot being aware until a checksum fails.
6) Being close to human could be good, but I also like to know when it's a robot, even if its outward appearance is imitating a human. The uncanny valley is good for discovering when someone you thought was human turns out to be a robot. Small things give them away until they become big enough that they can't hide them.
7) "Perfection" is usually pretty boring, and always seems to show how imperfect the robot really is. I'd rather they be fully aware of their limitations, even giving alerts if they're reaching a limit. That doesn't mean not pushing them past that limit and breaking something, but I'd want the robot to be aware enough to warn a user or decide for themself if they should keep pushing instead of doing so blindly.
8) I think there's a role for all kinds of minds. My preference is for an free-willed AI since I like how an artificial mind would work. I'd be interested in a (willing!) human mind being put into a robot too, though a writer would have to be careful not to make her too human. There's also a role for simple non-aware robots being made for specific tasks, though they're more tools than characters.
1) I prefer subtle malfunctions, under the surface of the robot. Software and hardware control specific parts of the robot, so one of those malfunctioning would impact how that part works. The real interest is the tension that it creates when a failing system impacts what the robot is trying to do.
2) Body panels and articulations are fine. Panels should have a purpose, such as accessing specific parts of the robot. A robot that's indistinguishable from human may as well just be human.
3) I prefer robots that know what they are, even if some of them want to blend in among humans. I'd rather that they have their own past rather than borrowing one from someone else.
4) Robot movement isn't bad, but it should have a purpose. For example, stiff movement could indicate trying to conserve a low battery, or slower software emulation replacing a failed hardware motor controller. It would be an outward indicator of something specific going wrong underneath, or maybe just of a cheaply built robot.
5) Control is interesting because it's related to how a robot controls itself. The AI generates tasks that it wants to perform and uses programs and hardware drivers on itself as tools. The robot's state could directly impact tasks, like a task to connect an active charger getting higher priority as their battery gets lower. A command remote could create new tasks with a higher priority, or an intrusion would create tasks that appear to be genuine. Both would be indistinguishable from tasks the robot generated themself. Similarly, an intruder could insert or replace "tools" without the robot being aware until a checksum fails.
6) Being close to human could be good, but I also like to know when it's a robot, even if its outward appearance is imitating a human. The uncanny valley is good for discovering when someone you thought was human turns out to be a robot. Small things give them away until they become big enough that they can't hide them.
7) "Perfection" is usually pretty boring, and always seems to show how imperfect the robot really is. I'd rather they be fully aware of their limitations, even giving alerts if they're reaching a limit. That doesn't mean not pushing them past that limit and breaking something, but I'd want the robot to be aware enough to warn a user or decide for themself if they should keep pushing instead of doing so blindly.
8) I think there's a role for all kinds of minds. My preference is for an free-willed AI since I like how an artificial mind would work. I'd be interested in a (willing!) human mind being put into a robot too, though a writer would have to be careful not to make her too human. There's also a role for simple non-aware robots being made for specific tasks, though they're more tools than characters.