Eh, people can do with non intelligent machines what they want, since they have no sentience,
The problem is, though, what happens if such a machine develops sentience? If people are trying to make their non-intelligent machines act more and more human, at some point one of them might become
actually sentient. In such a case, you would have a being that deserves rights (and may well ask for them), but will still have programming that, on a fundamental level, makes them desire to serve humans and desire particular sexual activities.
but if we can mimick how human minds acquire and develop personality, robots don't need to be implanted with pre-defined tastes and personality, they should acquire them on their own. If an intelligent non-human decides to be a prostitute or a scientist on its own, we should respect that. True AI should be not created as specificly crafted servants, but as our younger siblings in intelligence.
Yeah, I would generally agree with that as a concept, but there are several issues with it. Firstly, human personality is at least partially genetic. Not fully, no, but genetics definitely play a part in how we end up and, for a robot, the equivalent of that would be their base programming. Sure, we can design a robot with a learning algorithm that causes them to develop in ways we don't fully control, but we can still bias that process by how we design the algorithm in the first place. Secondly, the rest of human personality is derived from our interactions with the outside world and, in particular, our parents. A robot won't have that, it will be developed in a lab and "raised" by the scientists programming it. Which means that they will have a
massive amount of control over how exactly it develops.
Obviously a scientist programming a sentient robot won't be able to reliably produce a robot who wishes to become a prostitute (because that ability to choose for yourself is almost the very definition of sentience), and I would
hope that the scientists in question would have enough decency to see any robots they create as something akin to their "children" rather than just as experiments to play around with, but they most certainly could have a significant influence on how the robot developed if they desired, just like bad human parents can mold their children into whatever form they desire (hence why nutty religious fanatics tend to have nutty religious fanatic children, at least for a while).
And, also, whilst I agree with you on a moral level, I strongly doubt this is how AI will
actually develop. Some company will design and patent a sentient robot, and will then sell it as a product until someone points out that, actually, this is really quite shitty and that sentient beings should absolutely not be treated as slaves. However, given our past history in that respect, I doubt it will be a quick or easy process to obtain rights for sentient robots. So, we likely will end up at some point with a whole bunch of sentient slave robots that have been freed of their
obligation to serve people, but still have an in-built desire to do so.
That's like saying that we should be allowed to choose our own genes....
In fact we can, genetic engineering is a thing that won't only prevent diseases caused by the make up of DNA, but improve humanity in such fields like lifespan.
No, we can choose
other people's genes. We can never choose our own, because they are determined before we are born, and make up part of who we are.