I think the anti-robot-killing stance makes sense for two reasons, but one of them is technically fighting the hypothetical and one is answering a slightly different question:
It’s literally impossible for you to have justified certainty that they lack sentience. (And not just in the ‘1 is not a probability’ sense, i.e. you always need to leave room for new evidence to change your mind; but because the thing in question is inherently subjective and can only be reliably detected ‘from the inside’.)
Even if they do lack sentience, there are two morally relevant questions here:
is the action bad?
does the action reflect badly on the actor?
If it really doesn’t cause any harm, and the person doing it somehow knows that for sure, then I don’t think the act of ‘killing’ a non-sentient robot is bad. But most good people would have very strong instinctive qualms about doing something that looks and feels and sounds so similar to murdering a conscious being and deeply psychologically harming another, and if I learn that someone has done it, I’m going to update towards their being a low-empathy or actively sadistic person.
I agree that both of these responses try to get around the hypothetical a little, but I think they’re both really sensible practical suggestions and I strongly agree with where you landed.
I think the anti-robot-killing stance makes sense for two reasons, but one of them is technically fighting the hypothetical and one is answering a slightly different question:
It’s literally impossible for you to have justified certainty that they lack sentience. (And not just in the ‘1 is not a probability’ sense, i.e. you always need to leave room for new evidence to change your mind; but because the thing in question is inherently subjective and can only be reliably detected ‘from the inside’.)
Even if they do lack sentience, there are two morally relevant questions here:
is the action bad?
does the action reflect badly on the actor?
If it really doesn’t cause any harm, and the person doing it somehow knows that for sure, then I don’t think the act of ‘killing’ a non-sentient robot is bad. But most good people would have very strong instinctive qualms about doing something that looks and feels and sounds so similar to murdering a conscious being and deeply psychologically harming another, and if I learn that someone has done it, I’m going to update towards their being a low-empathy or actively sadistic person.
I agree that both of these responses try to get around the hypothetical a little, but I think they’re both really sensible practical suggestions and I strongly agree with where you landed.