AI and the Future of Book Covers

AI and the Future of Book Covers

by: T.L. Bodine

The internet being what it is, it really never feels like a good time to release a new book. You peek your head out like a groundhog to get a vibe-check of Twitter so you can make an announcement, and the coast is clear. Two seconds after you post, someone, somewhere, has been crowned the internet’s main character of the day, and your feed is awash in discourse. 

 

Or is that just me? 

 

One of the many hot-button recurring issues of the year so far has been generative AI, both the apps that churn content (ChatGPT) and the ones that spew images (Midjourney, Dall-E). Like a lot of things, this new tech emerged in a slow-building arc: 

1 – “Hey look at this new toy that does funny things!” 

2 – “Guys! Look how I can use this thing to make money!” 

3 – “Um, is anybody going to talk about how bad and exploitative this new thing is?”

4 – “People using this thing are both evil and uncool, away with them.” 

 

This journey should be fresh in everyone’s minds after the last few years of crypto frenzy. There’s another optional step five where the thing refuses to go away and instead it just becomes part of the fabric of life whether you want it to or not (see: my stubborn refusal to learn about TikTok). 

 

We’re in the thick of it now with generative AI, so there’s no way of knowing where it will ultimately end up, but among creative types – those writers and artists trying to earn a living by making things – the sentiment has settled pretty firmly on four. 

 

Except for the ones who are at two. 

 

Which has opened up a whole new dimension of perilous waters to navigate and consider as a strapped-for-cash creative gearing up to release a new book. In the simplest terms: On the one hand, creating eye-catching book covers just got a whole lot cheaper. On the other hand, using them might be career suicide. 

 

Let’s break it down. 

 

First Off, What’s the Big Deal?

 

The ethical debate over using AI generated art is a lot less about “what is art” and a lot more over the fact that AI generators are trained by digesting enormous quantities of artwork made by artists who didn’t consent to their work being used that way and who haven’t been compensated for it at all. Also, as it stands right now, AI-generated art can’t be legally copyrighted, so you really have no recourse if someone copies the entire piece and uses it for something else. 

 

I am firmly on the side of “art theft is bad” and “artists should be compensated.” I’m also sympathetic to the reality that there are people out there trying to make a go of earning money by selling book covers, and anything you can do to drop the overhead and production costs is probably mighty tempting. 

 

I don’t think it’s shocking that there are some designers out there using AI-generated artwork in place of stock photography, building compositions and manipulations around it the same they would with any other piece of licensed premade imagery. 

 

That certainly seems to be the case for art on Christopher Paolini’s book Fractal Noise, which earned Tor a great deal of criticism. Here, the art in question appears to have been generated elsewhere and placed for sale on a stock site, where it was then used by a designer in the final image. 

 

Despite some pretty loud backlash over the decision to use AI-generated stock imagery, Tor has forged ahead with the choice, and that may be the end of it…for them. For smaller creators, a controversy like that could be enough to ostracize you from the community you rely on. 

 

Doing Something Different

 

At pretty much the exact moment the internet was blowing up about the Tor cover debacle, my own book Neverest was facing a similar crisis. My publisher had just sent me a concept from the designer and I was vibrating with excitement. We were super close to being able to do a reveal when we discovered some of the assets used in the design were AI-generated. 

 

I don’t blame the original designer. Not really. I don’t feel like I have enough information about them, their circumstances, their art process, or anything else to be casting any stones. But it was clear to me and my editor alike that we were going to have to take a very different approach. 

 

 

Imagine a flurry of new concept mock-ups and reference collages being tossed back and forth as we tried to think of a new direction – not just a new artist, but a concept that absolutely could not leave us burned. 

 

We pivoted hard and reached out to Donnie Kirchner for something homespun. You can’t get more anti-AI than a physical painting that exists in space, and his work relies heavily on color and texture and lighting for its character. And unlike AI, which falls apart the longer you look at it, I think this pretty-simple painting gets infinitely more interesting the more you look at its depth and detail. 

 

I’m thrilled with the results. I’m relieved we were able to find something that would work. And I’m deeply sympathetic to anyone else who finds themselves facing a similar choice.

Tips for Indie Authors and Publishers 

 

Wherever you fall on the spectrum of ethics about AI, I think it’s a good practice to be mindful and at least know what you’re getting. Here are a few tips I can recommend for self-publishers and small presses looking to avoid using AI-generated art: 

 

1. Watch out for what stock images you use. You don’t have to generate it yourself to use AI-generated work. Just ask whoever designed the Paolini cover. Stolen art was already an issue plaguing some stock image sites, and now there’s an added complexity. Do your due diligence and reverse image search any piece of stock before you use it, and examine it more closely. 

• Know what to look for. Some AI-generated art is basically impossible to tell apart from the real deal. But there are a few tell-tale signs. Pay special attention to fingers and teeth. Look for shapes that don’t resolve and lighting that doesn’t make sense. Don’t be suckered in by shininess or overly processed renders. 

• Work with artists you know and trust. If budget allows, commission artists directly rather than buying from premade sites. Ask for samples. Pay attention to any difference in quality between samples or across their portfolio. And be suspicious if they are unable to provide edits, or if their edits suddenly drop in quality or consistency. 

• You get what you pay for. This isn’t always true, but it’s a good guideline in general that if you’re asking for something very complex, and someone is delivering it for a low price, either they’re taking shortcuts or you’re taking advantage of someone’s inexperience. Either way, if something feels too good to be true, it probably is. 

 

Will this AI situation lead to a resurgence in hand-painted cover art? I selfishly hope so, if only because – like Marge Simpson – I think they’re neat. But until we know for certain how everything will turn out, I say it’s probably best to err on the side of caution. 

 

ABOUT NEVEREST: 

One year ago, Sean Miller — journalist and mountain climbing enthusiast — reached the summit of Mount Everest and was never seen again. Unable to move on without knowing the truth of what happened, his widow Carrie insists on an expedition to search for Sean’s body so it can be properly laid to rest. Tom, Sean’s best friend and former climbing partner, agrees to serve as expedition guide and promises to keep Carrie safe on the mountain, despite their complicated relationship history.

Guided by a travel journal left behind by her husband, Carrie ventures into the frozen, open-air graveyard of the world’s tallest peak. But as Sean’s diary and Carrie’s experiences reveal, climbing the mountain is more than a test of endurance; it’s a battle of wills with an ancient and hostile force protecting the mountain — and the dead do not rest easy at the summit.