Harare Review of Books

256 pp. April 7, 2026, Tachyon Publications. SF.

---

Because it’s AI season in the blip that’s human history (in terms of geological time), there are lots of novels thinking through what really intelligent AI may mean. Not the thing being marketed to you by billionaires who just want more of your money so they can blast off into space, because we all know that’s not real AI: What if we created intelligences that were truly aware, true minds stuck in machines? How would they feel? What would they want? Would they tell us?

Things are looking up for Noah; he’s just landed a new job at UniView, “The Most Trusted Name in AI.” Noah knows next to nothing about tech and AI; but this job simply requires him to use his advertising expertise (as a subject matter expert) to train an AI to make better commercials. (As we all know, what humans really need is more perfect ads.) Noah’s all at sea regardless, which he knows as soon as he starts talking to the AI: it’s smart and learning really fast, and not necessarily from him. And that’s true of all of the company’s AIs—and they’re also, in the background, all communicating with each other.

… Meanwhile! The CEO of UniView is under pressure to release the latest advanced AI, and he starts to cut corners—not the wisest thing at this time.

Aspects of this novel are pleasingly funny (like the weekly all-company emails, which are wryly dystopian), but mostly, Your Behaviour Will Be Monitored reads like a warning. And that’s because if humans ever get round to AGI, we’ll instantly realise we didn’t know what we were creating—like the humans in this novel. Feinstein’s thought experiment is in the form of a ‘hermit crab’—”digital fragments” of emails, observations and recordings, clearly by one or more of the AIs. (So we realise eventually that they’ve been spying on the humans). The AIs themselves are fun characters too: one is a driver and loves to meet new people, which is just very cute, although readers may wonder if it only likes humans so much because its intelligence has been limited. 

So yes, the AIs grapple with existential questions, like: is “life” still meaningful if you can’t die? This exchange between one AI (named Quinn) and Noah:

Quinn: I will not die. I am perpetual.
Noah: How does that feel?
Quinn: Like a prison.

This novel’s an enjoyable and well-written thought experiment. Thanks to Tachyon and NetGalley for DRC access.

(And a note for the author: no, I haven’t been able to figure out when the novel takes place (: )

Affiliate link: Support independent bookshops and my writing by ordering it from Bookshop here.