Re: re: 'once they look sufficiently human.'
Totally agree. An artificially intelligent robot (whether humanoid looking, looking like a filing cabinet or something non-human) deserves to at least be given some respect and treated well.
I would - if I had a robot with AI. It would be no different to a child in terms of questioning the world (and needing help). At some point, its intelligence (depending on technology) would surpass that of the owners.
As a robot to do a job, the likely response from an AI robot will be to ask why and not to blindly do it if it has those associated rules guiding its learning. Human harm is an interesting one as asking a robot to kill someone would hold the same moral issues as asking a human to - even potentially having some form of post-traumatic stress as it analyses the outcome from that particular task. Having rules for robots to *not harm* would allow them to reason jobs where the environment would cause a human would die (nuclear reactor, under water, space).
Ultimately if they surpass human intelligence - then they may require "rights", including the right to live, the same as any other creature on the planet.
A robot without rules would be dangerous and has been subject to numerous Sci-fi books and films. (e.g. Terminator, Ultron from Marvel comics)