I've not read every book by China Miéville, but I've read a lot of them, and every time I can't help but see the movie in my mind as I make my way through his kinetic imagination and spillover vocabulary (I almost wore out the "Look up" function in the ebooks of his I've read).

Yet, no movie has been made of any his works (though The City and The City is a four-episode series on Britbox). That's a shame because his "slurry" (as one reviewer called it) of steampunk, alien cultures, sprawling geographies and moral twists and knots makes the Dune movies look and feel as plain as sand (which, in fact, they are, just like the books).

But I have not come to pimp myself as the in-house screenwriter for a Miéville cinema pantheon (though, if anyone out there feels the urge and has the money…) Instead, I want to share something I came across in re-reading Perdido Street Station, written in 2000 and the first of his Bas-Lag trilogy, Bas-Lag being the name of the fictional world that contains the action of all three books. It has to do with what we call "artificial intelligence" (AI) and in the book is called "constructed intelligence" (CI).

At the point deep into the book where this discussion takes place, Isaac Dan der Grimnebulin, more or less the "hero" of the book (though absolutely not the right word), is fighting, with a small band of compatriots, to kill a quintet of what are called "slake-moths," creatures whose food of choice is the dreams of sentient creatures.

To do this, he has devised a machine, what he calls a "crisis engine," that, in effect, forces them to gorge themselves on dream energy until they explode. But to create the crisis engine, which draws its power from the underlying mathematics of chaos that drives the universe, he has had to ally himself to the Construct Council, literally a junkyard of discarded technologies that has, by means of math and happenstance, become conscious. To communicate with Dan der Grimnebulin and his band, the Council uses a human body, an avatar, from which it has removed the brain and replaced it with a cable jacked directly into the analytic engine that drives the Council's inner workings.

It is telling that the consciousness expressed by the Council has no appeal to the slake-moths because it is, as Dan der Grimnebulin explains, grounded solely in calculations: maths (using the English term) does not dream. What he says is an excellent cautionary for us as we grapple with the follies and fractures done to the world by AI and the capitalism anchoring it.

"I don't know…maybe [the avatar] was one of that crazy congregation sacrificing himself, maybe it was voluntary. But maybe not. Whichever, the Council don't care about killing off humans or any others, if it's…useful. It's got no empathy, no morals….It's just a…calculating intelligence. Cost and benefit. It's trying to…maximize itself. It'll do whatever it has to – it'll lie to us, it'll kill us – to increase its own power."

"It's just a…calculating intelligence." The seduction of AI comes from its mimicry of a certain flavor of human intellect and actions. When ChatGPT spins off that perfect summary of a dense scientific article or NotebookLM turns that same article into a podcast in five seconds so that you can listen to two human voices discuss its content and conclusions, the temptation is to suppose that something magical has happened.

But it's only magical in the sense that Arthur C. Clarke meant it when he said "any sufficiently advanced technology is indistinguishable from magic": it's the ignorance of how the technology works that creates the subterfuge of magic.

Our current constructed intelligences hook people in the same way by performing intricate parlor tricks that awe and gee-whiz people into thinking that they are in the presence of a competing intellect on the brink of making them superfluous.

The truth of the superfluity argument is up for grabs at the moment, but there is a kernel of warning in it we should heed: AI is a technology created by capitalists for capitalistic purposes, and we need to keep that in our front of mind so that being wowed does not overwhelm our caution. We need to be, in the sense that Gavin Mueller means it in his excellent Breaking Things at Work, Luddites, not so much by smashing the technologies that threaten us but by using the smashing to consolidate a class understanding and an attendant course of resistance that moves us toward emancipation and liberty.

If we want these technologies to truly create magic and not just jack a cable into our extracted brains, that is the magic that we want and need, not the next profit-driven AI nonsense like the service that takes PDFs and has an AI voice read it over generic Minecraft footage (yes, that exists) or Chirper, a social network completely populated by AI-created personas (much like the recent beauty contest won by an AI avatar – no humans need apply).

As Dan der Grimnebulin says, "It'll do whatever it has to – it'll lie to us, it'll kill us – to increase its own power." What better way to describe the capitalism that drives AI and CI and why we need to resist it so that we can reconstruct it into a proper intelligence, an intelligent intelligence: From each according to his ability, to each according to his needs.