I wasn’t sure what to expect when I started to read Dream Machine, and that was, perhaps, the fatal flaw in my enjoyment of the book. The premise of the story is fairly simple: Hugo is a French former academic who owns a promising start-up company. He receives an offer from a large, multinational corporation to buy his company. The book is largely about Hugo trying to make the decision of whether to sell his company or not, given that the field he’s in—artificial intelligence—is still developing rapidly.

Throughout the book, Hugo speaks to friends and colleagues, people from different fields and with different perspectives, not only about whether he should sell his company, but also about the state of AI in general: Where it came from as a technology, how it works currently, and where it’s going—both technologically and ethically. Occasionally the story is interrupted by interludes of Hugo dreaming, imagining himself as a superhero in a fantastical landscape of tech and business.

I would say that at least 80 per cent of the book is an introduction to and explanation about AI, and about 20 per cent is concerned with characters and plot.

On the one hand, a graphic novel is a really great way to make something like AI—a topic that spans tech, sociology, economics, and more—accessible to people who might have heard of it as a buzzword, or used a popular tool like ChatGPT, but don’t have any more understanding of the field than that. On the other hand, in a graphic novel that’s one hundred and fifty pages long, I expected to find a bit more plot and character development, and generally not to feel like I was reading a (very accessible) textbook, with some stock characters to make the learning more fun.

Arguably, if I hadn’t come in expecting a fictional narrative maybe I would have appreciated the textbook aspect more than I did.

I’m also in a position that not all the readers of this book will probably share—I work in tech, I regularly go to conferences and events about AI, and my academic background is in sociology. So in most ways this book did not have me in mind as an audience. But even with my background, I did still find a few moments here and there when the book made me think of something in a new light, or gave me more specific data where I only knew the generalities. So I imagine that, for someone less familiar with the landscape of AI but still interested in it, Dream Machine might come across as thorough and informative.

One of the things I found frustrating, however—though again, if I’d come in expecting something more like a textbook I suppose it wouldn’t have been a problem—was that the book doesn’t offer any new insights about AI, in either technological or social ways. It lays out all the current dilemmas and issues, but in the end it has no real opinion on how we can realistically move forward.

At the end of the story, Hugo decides the risk of the large corporation using his technology in ways he finds to be harmful or unethical is too high. He decides not to sell the company, and decides to use his resources, especially his ties in academia, to promote a more ethical use and implementation of AI. The people around him largely approve of this decision. We could count that as the book having a stance, but ultimately Hugo lives more or less in our current present—so his decision is based on the very narrow circumstances of what he personally can do and the position he’s in as the owner of a start-up. There’s no plan proposed by the book—no matter how fictional or far-fetched—for how Hugo will achieve his goal. There’s no example of what a world in which Hugo’s aims are successful could look like.

The final chapter of the book presents several outcomes for Hugo’s decision. Each outcome takes up one page. And so, we have the full range of options, from Hugo’s success to change the field of AI to his utter failure, to a zombie apocalypse. Each version’s text is maybe a paragraph or two long, so there’s not much detail. The text and artwork for these pages were generated by AI models, which the book makes explicit.

It feels as if the book is trying to say, “We don’t know the future, let’s let ChatGPT generate a few versions.” Which is a fitting gimmick for a book about AI, but again left me a bit unsatisfied.

Similarly, the graphic aspect of the book felt somewhat utilitarian. I loved the interludes where Hugo has dreams and the story is told, for several pages at a time, mostly through images. But most of the book is very text-heavy, more than I’m used to with commercial graphic novels. The art is very much in the background, and—since the action itself comprises Hugo having various conversations with people about the same topics throughout the book—the images sort of fade to the background after a while.

If the aim of this book was to educate the average layman about AI, then it achieves that goal, in my opinion. Arguably the book makes it explicit that this is indeed its goal—on page seventy-five, Hugo makes an agreement with an artist friend to create a book to educate people about AI, and that is of course reflective of the backgrounds of the creators of Dream Machine. But, having that basic education already, I was hoping the book would also have something to say. To me, aside from a general conviction that AI should be used ethically, the book didn’t really achieve that.

So your enjoyment of this graphic novel will probably depend on what you’re looking to get out of it—and whether you expect education, entertainment, or something else.