I think you’re misrepresenting the way AI acts in the Culture. They’re really just depicted as hyper-competent humans, essentially, with their own goals, quirks

I think you’re misrepresenting the way AI acts in the Culture. They’re really just depicted as hyper-competent humans, essentially, with their own goals, quirks, and personalities. The reason The Culture works is because they’re self-policing: Banks depicts the AIs as having their own moral compasses, which do occasionally come into conflict, and in stories like Excession you see how the AIs form coalitions and deal with these conflicts. You also see AIs who don’t conform to their roles and are refitted, etc. Banks seems to believe in essentially universal moral principles that any sufficiently intelligent entity will have the capacity to reason out, and that there is a semi-inevitable progression towards a better understanding of and adherence to these principles as cultures develop (because cultures that don’t will destroy themselves). Or, at least, he believes that humans will successfully pass their own arbitrary moral principles on to AI; they are raised in our culture with our values, after all. The one thing that Banks does that I really like is that he highlights the fact that sufficiently advanced AI is not just a tool or machine the way we traditionally think of them in relation to ourselves; as thinking beings they are every bit as deserving of dignity as ourselves. Actually, the truth regarding the Culture is that the humans are more or less pets, in the sense that it’s the AI’s benevolence that allows the Culture to function at all. What makes this notion uplifting rather than depressing is that Banks sees a society of super-intelligences as inevitably benevolent, but not trivially so.