I write to explore artificial intelligence themes: How will strong AI emerge? What will AI culture look like? How will people and AI relate? How can we ensure that AI which is smarter, faster, and more powerful than people doesn’t eliminate or enslave us?
I’m certainly not the only writer to tackle these issues. One of the great joys is receiving emails from readers who recommend other fantastic books about artificial intelligence.
My goal here is to create a comprehensive list of the best fictional books about AI. (Because this could easily expand to nearly thousands of works, I’m focusing on novel-length fiction that reveals something significant about the themes mentioned above.) Overtime, I hope this will develop into a “Best of artificial intelligence” list.
From the list below, here are some of my top recommendations:
- Overall top five: Daemon, Computer One, The Lifecycle of Software Objects, and The Metamorphosis of Prime Intellect, and Accelerando.
- Where the focus is on plausibility of AI emergence: Daemon, The Lifecycle of Software Objects, and my own books: Avogadro Corp and A.I. Apocalypse.
- Hair-raising-on-the-back-of-your-neck scary: Daemon, Computer One, and Avogadro Corp.
Disclaimer: Links below, like elsewhere on my site, contain affiliate codes, so that I earn a few cents if you buy a book. Money earned here helps support my time spent on writing.
Without further ado:
Daemon by Daniel Suarez
I was about done with the first draft of Avogadro Corp and at work on editing it when I read Daemon, and nearly stopped writing. Daemon is mind-blowingly good, so much so that I was afraid I could add nothing further to the subgenre.
The basic idea is that a videogame designer dies, leaving his massively multiplayer online RPG running, with its AI set to take certain actions on his death. The AI has the ability to interact with the real world through text messages, emails, and phone calls.
Brilliant and scary.
Freedom (TM), the sequel, is a good book, but I felt Suarez took a pretty big turn in thematic direction between the two books. The sequel will differ significantly from your expectations at the end of the first novel.
Computer One by Warwick Collins
In Computer One, Warwick Collins lays out a compelling argument for why it’s likely that AI would try to preemptively wipe out humans. I think it’s an important read in the field of AI. It’s a fun read, and although some complain that it’s too philosophical or “talky”, if you enjoy novels by Cory Doctorow or The Lifecycle of Software Objects, you’ll enjoy Computer One.
Published in 1993, it was written earlier, and the computer references suggest it might have been written in the early 1980s or even later 1970s: the protagonist gets frequent computer printouts, and interacts with Computer One only via a “librarian.”
The Lifecycle of Software Objects by Ted Chiang
The Lifecycle of Software Objects is a wonderful story about how complex AI will grow and learn much the way humans do. I suspect that much of the early-generation strong AI will be like this, and we’ll end up with tech startups whose speciality will be training and educating AI. The book deals with the question of free will when the robots lose their research funding and have to become self-supporting. They still have limited intellect, and there’s a chance they’ll be abused by the power who hire them. What role do people have in protecting them, what choice do they have in what happens to them?
The Metamorphosis of Prime Intellect by Roger Williams
There is some brutal violence in this book (the first section reads like a snuff novel), and it’s not going to be for everyone. It is germane to the storyline though, and you just have to accept it for what it is.
Despite this, if you’re seriously into artificial intelligence fiction, this is a must-read.
The Metamorphosis of Prime Intellect is a brilliant thought experiment about the ramifications of a cornucopia world – where every wish can be granted – and the relationship between AI and humans in that environment.
I found the forest sequence to be less interesting than the main storyline (although reminiscent of one of my favorite novels: Earth Abides), but I can see how and why the author wanted to close up the story that way.
WWW Trilogy by Robert J. Sawyer
Published in 2009, WWW: Wake is the first book in this trilogy about Caitlin, a formerly blind teenager, who gets an experimental implant to allow her to see, and the emergence of an Internet-spanning global artificial consciousness.
Although it’s a fun story geared towards the young-adult reader, it’s low on the plausibility front (both in terms of the mechanism of emergence and the characterization of the AI.) However, Sawyer still tackles how girl, society, and AI come to terms with each other.
Accelerando by Charles Stross
Accelerando is the book that changed how I thought about the entire field of science fiction. Stross made it so that any science fiction novel that didn’t consider the technological singularity seemed implausible. Feline aside, even though the novel focused primarily on humans and post-human experience, the entire catalyst for change came from the AI.
Avogadro Corp by William Hertling
I might be biased since I wrote it, but I think Avogadro Corp is a very realistic AI emergence scenario. When I wrote it, I strove throughout to make the technology and human reactions as accurate as I could. One of my favorite pieces of feedback was from a reader who said that it gave her nightmares for days. I’m sorry for her discomfort, but I’m glad it felt that realistic.
This is the first of several scenarios I wanted to explore about emergence, and it’s the notion of early emergence, such that you can get only from the combine power of many millions of computers working together.
A.I. Apocalypse by William Hertling
A.I. Apocalypse, my second novel, was written to realistically explore what an AI culture would look like. How would AI interact? Judge each other? Make decisions? Although the AI themselves are anthropomorphized enough that I could make them into meaningful characters, the underlying system of communication, decision making, and economic trade is not.
In A.I. Apocalypse, I explore the emergence of strong artificial intelligence through the rapid evolution and mutation of a computer virus, followed by the development of a civilization composed of many hundreds of thousands of individual AI entities.
I haven’t read the following yet, although they are on my to-read list:
- When HARLIE Was One by David Gerrold (Codex recommended)
- Colussus by Dennis Felham Jones (Codex recommended)