„‚There’s nothing there.'“
Capaldi to „mother“, addressing the big question: Can a human being actually be perfectly replicated as a robot? He claims there is nothing that can’t be copied, nothing that would allow you to discern a perfect copy from the original. „Mother“ is doubtful about this, coming from a generation still carrying the „old feelings“, as Capaldi puts it.
Of course, this is a highly interesting question. Capaldi argues, that Klara becoming Josie would pass a higher-order Turing test (Turing test – Wikipedia ): Not even the mother can tell the difference between her child and a robot. Klara has been observing Josie for some time now, day and night. She has (probably) unfailing memory, so she could easily copy Josie’s behaviour in similar situations in the future. But she has no access to Josie’s genome (as far as we know), and a robot nourished by the sun will have difficulties to mimic the impact the microbiome has on human behaviour. And then also: Klara may be outstanding in monitoring and copying others, but does she fully understand the motivation of human behaviour? Will she be able to act like Josie in situations she has not been able to monitor in the past? So Capaldi’s claim seems to stand on a rather weak foundation.
There are also more issues with this: the premise entirely ignores the possibility that Klara herself might have been developing a personality, a self during her observations of humans, her life in the world. We hear her voice, we see her reaction to the pain of others, we also see her emotional connections to Rosa, Josie, Rick, the mother. She is a sentient being, even though she is objectified by the very nature of her position in the household – artificial friend, i.e., a friend that you can use when you want to and forget about when you have no need of her; a being who can be treated as a person when you need attention, and then left standing with her face turned to the fridge when you are too busy with other concerns/friends. We quickly forget how fraught with contradictions the very term is – it is conveniently abbreviated to AF precisely for that purpose.
Klara is in position of an object that is elevated to an almost-human status – given consciousness – and then demoted to the status of the object again, all the more cruelly, while she becomes more and more human through her observation and learning skills. And what Capaldi is proposing is a twisted form of slavery: she will not only have to submit all her outward actions to the demands of the human owning her, but actually kill her own inner self, adopting that of another. Turing test? That would mean a robot who proves her sentience and displays signs of consciousness indistinguishable from human is granted the same treatment as humans. But this is a Turing test twisted into torturous exploitation – a sentient robot is forced into an ethically dubious position (which in itself should be difficult to sustain for Klara with her inner purity) and asked to leave behind her individuality, to become someone else, with the assumption that her future (better) treatment will be based on her precisely NOT being herself, not acting on her conscious choices. She will be the mockery of a human – a simulation par excellence, with “nothing there”.
Yes – and on a more technical level, the question even is how such a „twisted form of slavery“ could be achieved: Klara has her memory, her history, or – as you point out – maybe what one could call her „personality“. She has involuntary flashbacks or associations. Is she actually in a position to „kill her own inner self“ if she is ordered to do so? And if yes: she needs access to data accumulated so far to emulate Josie – would these data still remain accessible for her?