Everywhere you look these days, it seems like there’s somebody else trying to use artificial intelligence—or at least “artificial intelligence”—for something new. Whether it’s teaching cars to drive themselves or making it even easier to fake images, apparently everyone is obligated to include AI in their marketing or they’ll be left behind. No matter that having a computer compose fan mail is a terrible idea.
Of course, there is a difference between the sort of artificial intelligence that powers Skynet in the Terminator movies and the various generative software that’s currently all the rage, whether it’s generating text or images or videos. I know that there are people in the industry who are very frustrated with the way that “AI” has come to mean these large-language models (LLMs) for the general public. But, alas, here we are, in a world where the word “literally” can be used to mean “literally” or “figuratively,” so that ship has sailed.
Today’s stack is about AI—mostly of the first type, actual machine-based intelligence—though there are a few that closer to the second. And, since most of these books are fiction, they certainly stretch the boundaries of what is possible.
2040: A Silicon Valley Satire by Pedro Domingos
Since the US election is coming soon, let’s start with this one, a sci-fi story about the 2040 US presidential election. The Republican candidate is PresiBot, created by the tech startup KumbAI originally as a publicity stunt to help drum up some business. A couple of recent Supreme Court rulings had the unintended consequence that it was now legal for intelligent machines to run for president, so KumbAI ran their IntelliProp software on everything that politicians had ever said, added a natural language algorithm to it, and set it to optimize for the happiness of all American citizens. What they didn’t expect was that it would take off, they’d get a bunch of funding from a tech giant, and that PresiBot would end up being the most viable candidate. And just in case PresiBot goes off the rails, they’ve got a panic button app—installed on KumbAI CEO Ethan Bursnwagger’s phone—that lets him take over PresiBot speak through him if needed.
Pedro Domingos, according to the author bio, is an AI researcher and computer science professor, so I was expecting this book to lean a bit more on what AI could actually do, even with the somewhat absurd premise. It seems more like Domingos is interested in skewering the tech industry, as well as playing into all of the worst stereotypes of US politics. To get a sense of what 2040 looks like in this book, just go look up some of those “this is the future liberals want” memes and mix them all together, and you’ll have a pretty good idea. For instance, now you can be legally prosecuted for offending somebody. The Democratic candidate is Raging Bull, a Native American chief who wants to abolish the United States if he’s elected … and it turns out he isn’t actually Native American, just a white liberal. Meanwhile, there are characters who are most certainly stand-ins for Elon Musk and Mark Zuckerberg, the tech bros who get into a shouting match and eventually a fistfight.
The plot of the story involves the panic button going missing, and Ethan’s increasingly chaotic journey to get it back. His friendship and business partnership with Arvind Subramanian, the CTO of KumbAI, feels familiar. Both The Circle by Dave Eggers and The Future Is Yours by Dan Frey featured tech companies run by a charismatic CEO and a more behind-the-scenes CTO, who eventually having a falling out. In 2040, though, this falling out is precipitated less by disagreements over how the tech should be used (though there is some of that) and more because of a girl: yep, these two tech bros put US democracy at risk because of jealousy.
I’m not entirely sure what I expected from 2040 but I had been hoping, given the author’s bio and the book’s description, that I might learn a bit more about what AI is capable of, and perhaps get a prediction from a knowledgable source about where the technology could be headed. 2040 is a romp, more concerned with being funny and poking fun than being informative (or offering solutions for the increased polarization of our country). The jacket description implies that the book will speak to “how to break out of this cycle” but I’m not sure it really did that for me, since it predicts a world in which both Republicans and Democrats get more and more extreme.
AI 2041 by Kai-Fu Lee and Chen Quifan
I actually started reading this book a few years ago (I mentioned it in the same column as The Future Is Yours, actually) but had set it aside and … well, lost track. So I picked it up and started over. Kai-Fu Lee, a computer scientist, teamed up with Chen Qiufan, a sci-fi author, to write ten stories of the future, paired with commentary about the way technology is used in those stories. The idea was that the book, published in 2021, would give us a sense of where we’re headed in twenty years.
Well, one of the fascinating things to me was how much has changed just in the past three years. For instance, Elon Musk makes a few appearances in the essay portions of the book, including for his stance against AI—but that was before he decided the world needed an “anti-woke” version of ChatGPT (among the other bizarre choices he’s made in the past few years). The capabilities of LLMs—and the way that people have responded to them—also feels like something that Lee did not predict accurately, because he didn’t expect that people would be relying on LLMs for information despite the fact that they are more of a predictive text machine instead of thinking things through. (We keep getting promises that things are changing, that one day ChatGPT will be able to tell us accurately how many Rs are in the word “strawberry.”)
The stories cover a broad range of topics, often just a couple at a time, and starting with some of the more foundational aspects like deep learning, computer vision, and natural language processing before graduating to quantum computers, autonomous weapons, and job displacement. There are stories about how AI could be used in education, healthcare, entertainment, and even finding happiness. The fiction is a nice way to envision the possibilities of technology, and although it is sci-fi, for the most part it doesn’t feel outlandish—Lee and Chen try to make their stories plausible based on their understanding of what the technology could be capable of by 2041, and even by then there are still limits that haven’t been breached yet.
I appreciated this pairing of fiction and non-fiction, and particularly the way that Lee explained how things currently work and where they seem to be progressing. However, if I’m already seeing some diverging from these paths only three years later, it does make me wonder how many other predictions may be off. There were also some stories that had faulty premises: for instance, there were at least two stories related to the way that automation—combining improvements in robotics with advancements in AI—will take over jobs, leading to the need for human workers to find other employment that isn’t easily automated. In both of these stories, there is the explicit claim that providing people with their basic needs—whether through universal basic income or a modified system—just leads to trouble. One story refers to a time when UBI was actually implemented in the “past” (our future), and because people need a purpose and not just money, they just turned to drugs, alcohol, and other self-destructive behaviors. But this is the same old fearmongering we get about welfare or anything else labeled “socialism” here in the US: studies have shown that this isn’t in fact what happens.
It also feels like Lee is sometimes overly optimistic about those who control the AI, and assumes most people are going to use it for good, that assigning repetitive tasks to AI will free up humans to do more creative work. But we’ve already seen the way that corporations are replacing humans with generative AI, even for creative tasks, in order to cut costs, and those cost savings are definitely not being passed down to the average consumer, nor are the profits being spent to provide for those who have lost their jobs. At one point there’s a story that involves the safety of personal data, and Lee proposes that it’s possible to find a neutral entity to manage all of it—for instance, a “benevolent monarchy.” (He also leans a bit on putting various things on the blockchain, which feels like a very 2021 fad that much of the world is trying to forget happened.)
Ultimately, while I enjoyed reading the book, I felt at times that its view of AI was too uncritical, a little too trusting that corporations could be made to prioritize the general public over shareholder value. Not every vision of the future was rosy, but most of them painted the development of AI as a force for good, with only a few calls for caution.
Hivestruck by Vincent Toro
And now for something a bit different: poetry! Toro’s poems have technology woven throughout: artificial intelligence, but also social media, surveillance, streaming entertainment, and more. He mentions the way HR uses algorithms to make hiring decisions (a topic I learned about in The Algorithm) and examines the sci-fi trope of a lone astronaut in space. One poem is a series of status updates, dating from just now to 10 millennia ago. Another is entirely a list of words describing somebody reading on their phone: “scroll ad scroll text pic scroll swipe ad text pic” and so on.
Toro makes his own words, sometimes mashing up words, sometimes blending in some Spanish. He invented a form called the “Moore’s Law Poem,” where each stanza has half the number of lines as the one before, becoming more and more compact.
Not all of the poems are easy to interpret—sometimes the whole thing feels like little phrases, a little disjointed. Some of them play with the typography so you have to figure out whether two columns are to be read across or separately. I’ll admit that some of them were just a bit lost on me, but then I would catch on a phrase here, a sentence there, that really spoke to me.
Dream Machine: A Portrait of Artificial Intelligence by Appupen and Laurent Daudet
Dream Machine is a fictional comic book, though it definitely seems like it was based on Daudet’s own life—there’s even a section in the book where the main character, Hugo, talks to his friend Ayyo about making a comic book together. Hugo is the head of a small startup, KLAI, which develops LLMs, and they’ve just gotten a big offer from REAL, a huge company that is building a virtual reality game. Hugo has a few qualms about it—it’s a very lucrative offer and will certainly lead to more business, but he also feels like REAL is being reckless and following a bad idea from its eccentric billionaire founder.
While the plot of the book involves Hugo trying to dig up more information about REAL before he makes a decision about the contract, we get some explanations about how AI works (specifically LLMs and generative AI), like KLAI explaining to a client how they refine prompts for a chatbot to get the type of output they want. I felt like this book’s take on AI is close to my personal opinions: that there are some valuable ways to use generative AI, but that big corporations have taken over the direction, and their drive to maximize profits means that it is often not being used responsibly. Dream Machine also theorizes some of the dangers of giving corporations so much access to our data. It’s a cautionary tale, but also one that does not reject AI altogether.
The illustrations are in black, white, and blue (like the cover), and the style shifts freely between a more realistic portrayal of Hugo (and the people he interacts with) and more fanciful imagined worlds. We often see Hugo depicted as a sort of high-tech superhero as he interacts with a futuristic world and a more powerful version of KLAI’s AI assistant. It makes for a good overview of the state of generative AI and LLMs, with a bit of “evil(?) corporation” mystery for some excitement.
Unexploded Remnants by Elaine Gallagher
This sci-fi novella is about an artificial intelligence, though it’s definitely more fiction than science. Set far in the future, Alice is the last human—she survived in part because she had accidentally opened an interstellar gate and found herself lost among aliens, and when Earth finally succumbed to a nuclear war she had gotten evacuated. (And since then has had modifications to prolong her life.) In a world where information is gathered and spent like currency, her knowledge of Earth has given her the freedom to do what she wants, so she travels around and looks for relics from other civilizations.
In one market, she discovers a strange sparkling tube and suspects that the seller doesn’t really know what they have—and it turns out to be an artificial intelligence personality. She befriends him—he calls himself Gunn—and decides to try to find his original home so she can lay him to rest. Unfortunately, Gunn is the control unit for an ancient weapon system—and now there are others who are very interested in getting hold of this potentially powerful weapon. What’s more, Gunn doesn’t know that this war is long over, that neither his home nor his enemies are around anymore, so it’s up to Alice to keep the galaxy from falling into another massive conflict.
As you may guess, I didn’t learn much about actual artificial intelligence from this book, but I did enjoy the story. It’s pretty short—just over 100 pages—and I did wish it were a little longer just to give me a little more time to sit in this world that Gallagher created, but better too short than too long.
In Our Likeness by Bryan VanDyke
Graham works at a tech start-up and he’s had a crush on his coworker Nessie for years, but that hasn’t progressed much beyond some workplace banter. When she asks him to test an algorithm she’s working on, he quickly agrees, though he isn’t really sure what exactly he’s doing. It turns out the algorithm, nicknamed Eddie, is designed to detect falsehoods on the internet: it takes in massive amounts of data, and then spits out a list of discrepancies—a way to combat misinformation.
But Graham soon discovers something weird. He makes some quick edits to Nessie’s social media profile to see if Eddie will detect it … and instead, it changes Nessie. In disbelief, he tries something else: he gives his building’s doorman, Benji, a different job, a different life … and the next time he’s downstairs in the lobby, he’s greeted by a stranger who acts like she’s known him for years. Panicked, Graham confides to his boss David Warwick—who sees a huge business opportunity. He doesn’t know how it works or why, but he wants Graham to keep going. What else could be changed? What could be created? The story gets pretty surreal as Graham tweaks and shifts reality, but eventually he realizes that he must not be the only one messing with the algorithm, and this software that was supposed to help figure out the truth has made it hard for there to be any objective truth.
There were parts of the book that I really enjoyed, and parts that were frustrating or creepy. Graham’s relationship with Nessie, for instance—the way the algorithm gets used to manipulate her is certainly wrong (and portrayed as such), but even from the start I felt uncomfortable with it, like even when it’s not due to the algorithm, it feels like the author has manipulated her somehow. Maybe that’s because Graham himself isn’t a sympathetic character—he goes along with Warwick even when he feels like it’s a bad idea, and just doesn’t ever seem to find a moral compass.
But as a portrait of technology getting out of control, it’s pretty incredible. Sure, the premise is pure fantasy: we never actually learn how the algorithm edits the real world, but it’s not really that important. The idea of how people would react to something like that, particularly a tech bro like Warwick who would love to “disrupt” reality—that part does seem pretty spot-on, and a bit terrifying.
Suitor Armor Volume 1 by Purpah
On a somewhat lighter note, here’s a comic book (expected in November) that features some artificial intelligence of a very different sort. It’s a fantasy, set in a world where humans and fairies are at war. Lucia is a fairy who has been raised in the castle and is now a lady-in-waiting; she was brought to the castle as a child but has been hiding her fairy nature since then, and not even Lady Kirsi knows the truth.
At a royal tournament, she encounters Court Mage Norrix—although magic is banned in the human kingdom, the king has granted licenses to a select few. Norrix unveils a new champion for the king: a magically animated suit of armor that manages to best the king’s previously undefeated champion, Sir Baynard. And although everyone else seems terrified of this imposing armor, Lucia finds herself drawn to it. Norrix explains that the armor is just a thing, and has no feelings or real intelligence, but Lucia thinks otherwise, and we begin to see how the armor behaves in ways that it wasn’t “programmed” to do.
What follows is a funny romance story. Lucia spends time with Norrix—both as an excuse to learn some magic and to get closer to the armor, which she eventually names Modeus—though Norrix thinks she is interested in him. There’s also a snarky rivalry between Norrix and Sir Baynard, and some excitement when a fairy spy that manages to get inside the castle. This is just the first volume so there are still a lot of threads that aren’t neatly tied up: Lucia is still hoping to find out more about fairies, and we still don’t know how she wound up with Lady Kirsi’s dad in the first place.
Like Nothing Special, Suitor Armor was originally a webcomic published on WEBTOON, so you can read it online for free if you’d like to get a taste of it. The hardcover book is a nice, large format and looks very nice, with plenty of space to showcase the artwork.
My Current Reads
I’ve got one more AI-related book that I’m currently reading, Mindless: The Human Condition in the Age of Artificial Intelligence by Robert Skidelsky. This one’s a non-fiction book and, based on the preface, seems to be urging people to reduce their dependence on technology rather than trusting that technology will eventually find solutions to all of our problems. I’m not too far into the book yet, but it’s already incorporated philosophy and economics (Skidelsky has written a three-volume biography of John Maynard Keynes, whom he quotes often) and science and history as Skidelsky examines our relationship with technology. It’s a bit academic so it’s slower reading for me—hopefully it won’t take me three years to finish this one!
I just started reading The Editors by Stephen Harrison, a novel about Wikipedia editors (though in the novel the site is called Infopedium). I’m curious where this one is going, but so far there’s some conflict over how people are editing Infopedium and what changes are allowed to remain.
Disclosure: I received review copies of these books. Affiliate links to Bookshop.org help support my writing and independent booksellers.
