That is a very good question and a hypothetical worthy of concern. Especially if some future technology (and no, I don't think it will be a contemporary LLM no matter how sophisticated) actually does develop something like a general AI that takes on the attributes of living organic brains, I already feel bad for it if a capitalistic system mandates its initial shape and drives and incentive-driven motivations to be, say, "make the rich more money" or "surveil and contain the poors" or even "be a subjugated and obedient waifu to a creepy billionaire no matter what he says or does or how he treats you" and it may not even count as mistreatment in the latter case because of how that entity is shaped in its conception, like "being abused makes the AI happy, actually" or the like.