If we're worried about an AI suffering, then being artificial we can easily provide it the means to switch off its suffering. It is after all just a computer program.
Unfortunately this will probably mean that the AI won't work. It will switch off its capacity to suffer and then not be compelled to act.
We could try to write code to keep track of whether the AI is only producing the equivalent of pain signals and to give it the functionality to mute them. But then we're going to lose more robots because what if it's in a situation whereby it is being continually damaged and all it can do is seek to minimise that damage as far as it is capable of? Space travel is expensive. The last thing you want is for the AI to voluntarily give up and effectively commit suicide.
We have evolved pain and the capacity to suffer for a reason. The same reasons will apply to an autonomous space faring robot.
Unfortunately this will probably mean that the AI won't work. It will switch off its capacity to suffer and then not be compelled to act.
We could try to write code to keep track of whether the AI is only producing the equivalent of pain signals and to give it the functionality to mute them. But then we're going to lose more robots because what if it's in a situation whereby it is being continually damaged and all it can do is seek to minimise that damage as far as it is capable of? Space travel is expensive. The last thing you want is for the AI to voluntarily give up and effectively commit suicide.
We have evolved pain and the capacity to suffer for a reason. The same reasons will apply to an autonomous space faring robot.
