Peter M. Ball

After a few years of ignoring AI-based art generators, a few of the major players finally made a move that made them really interesting for me: they announced that the product of the digital minds could be licensed for commercial use.

This happened right as I was playing around with cover designs for a back list release, so I figured it was as good a time as any to see if an AI could generate a passable cover for one of the tricker books to get right with stock art images. The image above is my first real attempt, after playing around with some concepts and art styles. 

They’re an imperfect tool, both because the art generated is basically a crapshoot and because the software typically has some prohibited terms. Asking the mad robot artist for a unicorn scull on a desk surrounded by bullets or revolver was prohbited, for example, although asking for a painting using the key words unicorn skull, detectives cluttered desk, noir ended up with something usable after a little time in photoshop.

This isn’t a tool I’d necessarily turn to for a major release , particular for Brain Jar, as it’s neither accurate enough nor…well, at my level, pretty enough to make for a great cover. But not everything is a major release – sometimes, books get put out *in case* someone wants to buy them in a particular format. There’s already perfectly viable singles copies of both Horn and Bleed, the two Miriam Aster novellas, out there with some pretty damn gorgeous cover art commissioned by the original publisher.

What doesn’t exist – and what I hold the rights to produce – is a book that collects them together. And I suspect the above will be enough for folks who may be interested. I’m still pondering it for another few days, while I dig into the minutia of the AI system I’m building with, but it’s definitely got some potential. 

On the other hand, AI software is still an untest territory so far as copyright is concerned. Just beacuse the software developers are granting commercial use rights, it doesn’t mean that its clear terrain. Copyright law has a lot of precedents around art needing to be generated by human intelligences – something that’s come up in elephant generated art and photographs taken by monkeys – and early test cases are suggesting AI creations aren’t copywritable. 

Similarly, there’s the ethical side of things: AI software is trained by feeding countless images and artists work into the learnign program and teaching it to make connections, and there’s still plenty of concerns – both legal and ethical – about how those things are spat out the other side.

So while I’m playing here, it’s not a guarentee that it’s the cover I’ll go with in the end (although its’ defintely cheaper and faster than my next option).

For the curious, some of the other AI creations it coughed up while I was playing with Miriam Aster images are included below. 

When talking back list, one of my core lessons is you have time to capture the reader. Often that means you can sell them a book months or years after they first encounter you, and if you do things right, you can sell them books at full price instead of constantly discounting.

Here’s a little case study of what that looks like, drawn from a recent purchase. 

PHASE ONE: THE LEAD

I first encountered the Phillipine’s based romance author Mina Esguerra while working for BWF a few months back. Their work had an interesting hook that captured me immediately (Esguerra is the founder of Romance Class, a grass-roots collaborative movement that’s basically built up the English language romance scene in the Philipine’s from scratch).

This was basically a lead event — I’d encountered the author, discovered some interest, but didn’t immediately want to go out and read their books. I was more interested in what Esguerra did than what they wrote.

But it did get me to a point of engagement — I attended their session at BWF — and convinced me to keep track of their work via their twitter feed.

PHASE TWO: WARMING THE READER

Long-time readers may remember that I loved their BWF session, which certainly stoked my interest in reading their work, but didn’t quite tip me over into buying a copy. Not least because funds were short at the time — I was in between my last BWF paycheque and my first paycheque from the writers centre — and romance is a genre I love, but definitely not my core reading genre.

In marketing terms, I was a warm customer. Interested, but not yet convinced the value of Esguerra’s fiction would be worth the opportunity cost of the purchase price and the time spent reading.

Readers often spend a lot of time in this phase. We’re distantly aware of countless books we might enjoy, but it’s not until there’s a surge of conversation we want to be part of, or great reviews, or personal recommendations from trusted friends that we finally break down and purchase.

Traditional publishing is great at building that value quickly, but it’s not the only way. Lots of indie author simply warm readers up gradually, putting out stuff folks are kind of interested in and talkinga bout interesting stuff on their socials, never quite making a sale but not doing anything to cut a kinda-interested reader off.

This process of warming a reader up is a slow, incremental increase in the potential value of a book until the opportunity cost ceases to be a barrier. There’s a lot of ways to do it ; short-term discounts can rewrite everyone’s opportunity cost calcuations, for instance, but it’s just as likely you’ll eventually put something out that hits the sweet spot for a reader’s interest.

Which leads me to…

PHASE THREE: THE FIRST PURCHASE

What finally tipped me over the line to purchasing my first Esguerra book? This tweet:

What tipped the value proposition in Esguerra’s favour wasn’t a price point, but the framing of these older works in terms that appealed to my interests. I’m very interested in series works and self-published short story models, and I’m a big fan of heist films and TV shows like leverage.

The book cost me $5.40 Australian, which is line with Esguerra’s other works, but it ceased being an obstacle because it hit a sweet spot in terms of value. Even if I dislike Esguerra’s writing, my $5.40 will not be wasted because there’s other sources of value in the book beyond the story.

It also coincided with a recent payday from work, and the arrival of my tax return, so I’m I’ve got a bit of a buffer in my impulse buys spending account. I’m primed to go from a warm lead (interested) to hot (paying money).

Total time to make the first sale from the lead time? Approxiamtely 12 months, although I’d been bubbling along with a vague interest that entire time. It was just the right combination of factors, many of which were outside of the author’s control, that finally tipped me over.

And, having tipped over and bought the book, I have a different data set to use when calculating in Esguerra’s other work. If I like what I’ve read, the next phase (superhot) could well be purchasing a lot more of her books, with much less time and effort required from her end.