I think the most interesting part for me is to analyze a little how this book positions AI compared to other books in the genre.
I don't know if this is just my own bias (and aversion to techbros), but there's something about this version of AI in this book that seems like it's meant to be an extension of LLM airquotes AI hype. It's something about the wording of "when AI began participating in creative tasks" that twigs that feeling for me. There's almost nothing here about sentience or consciousness. AI is both simultaneously a critical focus of the book while also relegated to the background.
We're told that AI has "worked magic on municipal planning and administration freeing the world of poverty" (citation needed). The book mentions one-off forensic evidence of "companies [that] were created by AI that emerged organically from global networks" but that's the last we hear of it. It also tells us that "no matter how we try to prevent it, corporations like LK are destined to become one giant AI" going further to say "we're all fated to disappear into the bowels of an AI behemoth" and this is arguably the conclusion of the book. But this is all talk.
The reader doesn't get to see any of this AI machination on page and so it's largely just an existential threat. How many AIs are there? Do they talk to each other? Even when characters are threatening to or partially merging with AI, we just hear about it, and don't learn what this means in the context of this world or for these characters. When the narrator interacts with AI, it's less about talking to a person and more like just using some tool: "external affairs AI lets me know where it thinks they're headed." That's the most detailed human computer interaction this book has to offer.
In thinking about this, here's a bunch of other books that came to mind that deal with these topics better:
Counterweight talks about the old president distributing their consciousness elsewhere, but what this means and its implications are barely touched upon, whereas that topic is a focal point for Ann Leckie's Ancillary Justice. The AIs in Counterweight are barely treated as tools and yet they are creating companies and merging with humans and exploring space in the epilogue, but there's no consideration of their sentience or consciousness or their rights in the same way that Ancillary Mercy does either.
Ursula Whitcher's North Continent Ribbon tells a much more nuanced story about the ways in which humans push back against machine intelligence and the messy line between human and machine responsibility.
There's a throwaway line of AI having its own sense of humor in picking music that humans don't understand, but Sue Burke's Dual Memory provides a much more nuanced and direct view of machine culture (as well more of a variety of what machine sentience and machine goals could look like).
I also think Charles Stross's Accelerando does a better job at describing the effects of powerful AIs that are so strange as to seem alien, and that have left no room for humanity.