Why AI and Publishing are a Bad Mix

(Disclaimer: I am not an expert on AI and while I have attached links to show where my information comes from, I am not accusing anyone in particular and am open to being proven wrong if further information comes to light. This article is purely my opinion on the current state of the publishing industry.)

A while back, my publisher Rhiza Press released a statement on social media regarding their stance on AI. They have made it clear that any projects (literary or illustrations) that have used AI in any form (including brainstorming) are not to be submitted to them for consideration. And I stand with them firmly on this.

It makes me proud to know I'm with a publisher who respects human creativity, especially in a time when bigger publishers like Harper Collins and Hachette have gotten themselves into big messes regarding books that are suspected to feature heavy AI use – most notably The Most Dangerous Games by James Patterson and Shy Girl by Mia Ballard (see link: Hachette pulls horror novel Shy Girl after suspected AI use | Books | The Guardian). It's uplifting and yet also a tad bittersweet to see that a small press from Queensland, Australia, is capable of upholding what should be industry standard, while these big companies shove books out the door without checking them properly, then hide the evidence under the rug.

But for those of you who might be asking, I thought I'd get deeper into the why of it all.

Why can't AI be used in publishing? Great question.

Well, one thing I clocked pretty early on was that this thing we call "artificial intelligence" isn't particularly ... intelligent.

The apps we know as AI are essentially very complex algorithms. That's why they are seen as useful in medical research fields, because they are a tool used to detect patterns among large swathes of data, something that would take a human a lot longer to accomplish.

But creative fields like art, music, writing and more have never been about algorithms. They were born from human experience, understanding, imagination and creativity. 1984 by George Orwell (a perfect allegory for this very conversation) was inspired by Orwell's own concerns regarding the trajectory of politics and surveillance, and his fear of a totalitarian regime. AI could never produce such a feat because it doesn't experience fear the way we do. It has no idea what that feels like.

If you'd like some further explanation on how AI works when generating writing, I recommend this YouTube video from Meredith Novaco. She has a degree in computer science and helpfully breaks down how an LLM works, far better than I could explain it (see link: what's with the Powerless AI allegations?).

But the real purpose of this blog post was to serve as a reminder – AI is a people pleaser. It is designed to spit out a response to the question, but who decides if the answer is satisfactory?

We do.

Humans have always been the decider of what is right and wrong, good or bad, correct or incorrect, and ethical or unethical. But it takes skill and knowledge to truly validate this. I find the best example of this to be crochet – a craft that, unlike knitting, can only be done by human hands. AI can generate a crochet pattern, but a human has to make it to determine if the pattern actually works (spoilers, they never work). The same can be said for writing. AI can generate a short story, but I, the reader, get to decide if it’s good or not. The same, of course, can be said for any human-written story, but at least in that case, I can respect the time and effort taken to write it.

Even Douglas Adams seemed to recognise this in The Hitchhiker’s Guide to the Galaxy (another book that AI could never have written, simply due to its brilliant absurdity and intentional nonsensicalness). In the novel, a supercomputer called Deep Thought is built for one purpose – determining the answer to Life, the Universe and Everything. After thinking hard for several millennia, it spits out an answer: 42. But it takes Arthur Dent, a human, to find the question that best suits this answer.

The problem we’re seeing in publishing is that the people who truly believe AI does have a place in the arts are the people who never had a place in them to begin with. Art of any kind is a skill. It must be learned and earned. Practice takes time and some people just don’t want to put in that effort. If that’s the case, then maybe the arts just aren’t for those people. And that’s okay. There’s plenty of other things to be good at. I myself will never be a mathematician or a professional basketballer, and I’ve accepted that that is my lot in life. But writing is something I’ve loved since childhood and dedicated hundreds of thousands of hours to. I studied it at university, put myself in debt over it, cried myself to sleep with every rejection I received. Writing is my whole heart and soul and THAT is why I have the ability to validate if writing is good or not (within reason, of course, art will always be subjective).

So it should come as no surprise that to someone who can’t write, AI seems like a literary genius. But to those of us who can, we see the flaws in its work immediately. We know it can’t write, we know it produces similes that make no sense, or boring, repetitive descriptions that often, also, don’t make sense. Because AI has no imagination; it doesn’t bear the same passion and devotion that real artists can’t let go of. It doesn't need to write to cope with anxiety, or grief, or trauma the way some authors do. That kind of pain is what drives us artists to keep going. It's not enough for us to just type in a new prompt and hit ‘enter’.

As for the publishers, they’re on a different side of it. They enjoy books same as the rest of us, but not all publishers are writers. At the end of the day, they’re a business and businesses need to make money. But when money becomes more important than talent or quality, it’s not just the writers who suffer but also the consumers. If you pay $25 for a book, you deserve a novel that has the bare minimum of good quality – namely, a structure that is logical, a plot that makes sense, characters that are interesting and prose that is readable at the very least. Unfortunately, there has been an increasing trend in traditional publishing of late with books being rushed out the door, poorly written, poorly edited, yet dressed up in gorgeous artwork and sprayed edges to distract you from the real problem.

It’s bad enough already, and bringing AI into it only makes things worse. I haven’t even begun to address the legal issues of AI. Generative AI is essentially theft and, in fact, cannot be copyrighted, therefore no publisher should be taking AI works because if you can’t copyright a work, you cannot make money from selling its rights. And for the consumers, AI in publishing means even worse quality of prose, plot and structure, and that simply isn’t worth your hard-earned dollars.

As for the writers, it’s an insult to our craft. Why are small writers being paid pittances for work they created themselves, when authors like Jerry Falade are being offered $2m deals only for it to be revoked after suspected AI use (see link: $2m crime novel deal collapses amid questions over AI use | Books | The Guardian). How does a novel make it through so many read-throughs and edits before anyone even notices? And why are publishers only listening when the novel is put through an AI detector – a tool known for being completely unreliable at detecting AI? Surely, an editor – someone qualified in the art of editing manuscripts – should be capable of picking up on the inconsistencies, the repetitive prose, the dry characters or dropped-off plot points. If readers can see it, then why can't editors?

In contrast, however, we are also seeing authors constantly accused of AI use, simply because a reader doesn't like what they are reading. Some of these works come from authors who have been openly and staunchly against AI use from the beginning and so the accusation is most likely unfounded. And yet, the witch hunts continue as readers become more and more paranoid about the potential presence of AI in what they are reading. And as infuriating as it may be from the writer's perspective, for consumers, it is understandable.

At the end of the day, it comes back to one simple question: if you can't be bothered to write it yourself, why should I bother to read it? The few authors out there who are openly for AI use in writing will learn in time the consequences of this decision as readers drop off and move on to other authors who are more passionate about their work. I can only hope that publishers do the same. Offering contracts to authors whose works are suspected of AI use or even generating AI books under a new author name (see this video on Silver Elite by Dani Francis: Silver Elite | AI INVESTIGATION), should be completely unthinkable. And yet, the unthinkable IS happening right under our noses. Unfortunately, without proper proof, we can only ever speculate and while speculation is enough to raise eyebrows, it isn't enough to stop it from continuing.

The reality is, using AI in creative spaces simply isn't a sustainable – or ethical – practice. Readers want heart and soul, they want emotion and effort – things AI simply cannot provide. Moreover, readers want to purchase from authors and publishers they can trust and the use of AI in publishing is diminishing that trust each day. And in an industry where marginalised voices are already struggling to find their space, it is an insult to them especially to make room for a machine, when real human beings are fighting to earn a seat at the table.

To put it simply, we – the writers and the readers – deserve better.