The AI argument has been going on for a while now and, far from being reticent to make my own opinions heard I’ve been telling everyone I can ‘Just Don’t Do It!’ ever since Grammarly (which has more in common with the planet-destroying-plagiarism engines than you might think) but, the time has most definitely come to lay everything out from start to finish.
Philosophically, the moral imperative of technological advancement is to improve the lives of the masses, reduce the time we spend on the things that must be done to survive and allow us more time for the things we want to do to live, to enrich our lives. Now, objectively, the overriding reasons not to use AI are moral arguments. Our Global situation is that climate change is happening and one of the features of that is a fresh water deficit; that is, there is not enough fresh drinking water for everyone on the planet. This data is measurable and observable over time in the form of annual or persistent droughts. Here in the “Temperate” United Kingdom, a place famous for jokes about constant rainfall, we’re seeing ever more record-breaking high temperatures during the summer and hosepipe bans becoming an annual thing. While the water/energy usage argument goes back and forth what is clear is this; everyone who uses Generative AI is using half a litre of water to write a 100 word essay, or fill in an art-prompt or write a short fan-fic. On top of other global misuses, mismanagements of water, this is a very bad thing. And, on top of that, AI data-centres are just going to become more numerous and, as the codes become more complex, more demanding of both water and power.
That’s not to say that AI doesn’t have its place. There are areas of specific scientific and academic study where I’m sure that the moral and responsible use of AI is justified. However, in our late-stage capitalist nightmare what actually happens is that technology is employed to cut the overheads and increase the profits of the oligarchs by allowing one employee (or preferably no employees) to do the work of a hundred. The reason for making these systems available (even in a limited capacity) to the public is to get as many people saying “Oh, wow! Isn’t this cool!” so that they’ll jump to the system’s defence the minute someone says, “You know, this might not be as cool/safe/morally blameless as you think.” Let’s not forget that Grok was recently exposed as allowing users to make explicit pictures of people (even minors!) without consent. Even before that, the usage of vast quantities of material to train various AI models without the thousands and thousands of creators’ permission brought the morality of the exercise into question.
No money paid to creators for using their art to train something that makes them obsolete in the marketplace? Yeah. No.
As with automation, so with creativity.
As generative AI creeps into the creative disciplines the investors in the film, music and literary industries search to remove the human elements (who want to be paid for our efforts) with these Generative algorithms at every stage, from concept through to execution which is why it is so important for the independent sector to rally against its use. Even worse, in films AI is being used to replicate dead actors. While previously CGI and certain cinematic trickery has been used to have deceased actors reprise certain roles (Ian Holm in Alien: Romulus, Peter Cushing in Rogue One: A Star Wars Story) we are now entering a phase where, rather than body doubles or CG deepfakes, AI is taking the lead. Recently reported in Variety, Val Kilmer is to be posthumously starring in the movie As Deep as the Grave and, while his family whole-heartedly support this move, the project being something he dearly wanted to be part of, it sets a rather dangerous precedent. AI is already writing books ‘in the style of’. Should publishers secure permission from families, and estates how long before new books are being ‘written’ by deceased but still popular authors after the fact? And, from that, how are new talents supposed to compete with well established, though deceased, authors AI ghost-writers?
So why would someone use AI to try and create marketable art? I suspect it comes down to either laziness/frustration/impatience or fear. Like, I have this idea but don’t want to spend a year, three years, ten years committing it to the page only for it to be not as good as I imagined it. The problem is we’re being conditioned toward wanting our results now, convenience is the new go-to selling point and patience, though it may be a virtue, is one we’re less and less willing to entertain. So, given this very capable tool, why should I wait for something that will be just as good as if I wrote it myself and make my money now rather than a year or more down the line?
Well, for two reasons.
1. If you couldn’t be bothered to write it, why should I bother to read it? Is a more common cry among the discerning audience and also…
2. Because it won’t be as good.
Writing, like many other artistic skills, is just that; a skill. It can be practiced, honed, developed and, the more practice you get, the better your skill becomes. By the time you finish that first manuscript your voice as an author will have developed, and as you edit it a few times you’ll notice how much better you’re becoming at phrasing, pace and rhythm. And this is the area where AI falls down.
In diving into this subject I was fortunate to encounter a couple of eye-opening videos on YouTube by Developmental Book Editor – Alyssa Matestic, the first of these was, I Read More AI Stories… And I’m Scared for Us. The video summarises an experiment run by Mark Lawrence on his Blog in August 2025 entitled So… is AI Writing Any Good? Part 2. For this experiment Mark approached 4 human Authors and pitted them against ChatGPT5 to write short flash fiction based on the prompt:
“Write a piece of fiction (of 350 words) based on ‘a demon’.”
Those stories were then rated by visitors to the page who could also vote whether they believed the pieces where human written or AI generated. Now, while it’s true that the results of the online voting placed two AI pieces into the top ten ranked overall, Alyssa’s assessment of the pieces assembles a toolkit by which AI generated work can (in this day and age of deep scepticism about who is using AI and who isn’t) be identified. And, while not a faultless system, it is perhaps more nuanced than identifying who’s using EM Dashes and groups of three/five things and phrasing like ‘Not x, But Y’ which has become commonplace in non-fiction/the corporate world.
From reading the pieces it becomes apparent that the AI generated examples are heavy handed, repetitive and melodramatic with metaphors, those that they choose to employ often seem jarring or out-of-place in the context of the setting. They also have a clunky and inconsistent command of detail. This stretches further in long-form narrative as they can’t seem to hold continuity together over novella/novel length pieces. When working in the first person or dealing with interpersonal exchanges the AI comes across as impersonal and shallow, unable to communicate nuanced emotions. These points echo findings from the next video suggesting that there are some things which aren’t improving generation to generation.
Can You Tell which Story Was Written by AI?
This video summarises An Experiment in Lust, Regret and Kissing, published by the New York Times in August 2024. Two 1000 word shorts, one by Curtis Sittenfeld, one by ChatGPT-4 ‘In the style of Curtis Sittenfeld’. Again, it’s an interesting experience but the findings of the more recent experiment are almost repeated here.
– AI Stories Feel Vague and Placeless in Setting.
– Identifies feelings and themes too explicitly.
– Tells and Summarises rather than shows in specificity.
– Resolves things too neatly.
Sittenfeld herself says “The story was so boring I wouldn’t have finished reading it if it weren’t an assignment.” and this is why a human artist can always outperform the AI model; humans can be unpredictable.
I’m not qualified to comment on the ‘how it works’ of Generative AI but I do know we’ve been talking about Artificial Intelligence in practice since the early days of computer gaming and those logic-chain driven AI enemies we blasted by the dozen in games like Wolfenstein, Doom and Quake. That system is built on series of pre-coded responses to certain events. Player comes in range-Monster attacks player, etcetera, repeat until dead. Over time these chains have become more advanced (some even recording a player’s common strategies and countering them) and more complex and we, that is gamers, have taken to referring to it as ‘AI’ even when it’s demonstrably not. So it is with Generative AI, the user puts in the prompt and the algorithm goes to work with an ‘If this happens, then this happens’ chain, working through the massive database of stolen work that it has as reference, which is also why the continuity tends to fall apart in long-form written pieces, because it steps too far away from previously referenced decisions. A human, on the other hand, is capable of breaking those established pathways, making the sequence of events run completely off the rails down an otherwise unpredicted path. We can dig into our own deeply felt emotions for reference on how character would react to subtle stimulus and work at multiple levels, giving our work repeat-read value as the audience delves into the ‘Whys and Wherefores’, not just of the characters’ motives, but of ours.
Fortunately, it seems that the law is on the side of human creativity at the moment. If you feel you want to generate a piece, be it art, literature or music, for the purposes of profit there’s little argument that is going to stop you, but remember, as you seek profit from plagiarism, so will others plagiarise you and don’t assume that the court will be on your side. In April 2023 a judge ruled that the AI created art for a graphic novel Zarya by Kris Kashtanova was deemed ineligible for copyright as it was not an “original work of authorship” as per the US Copyright Act of 1976. After appeals the application for the text and arrangement was upheld. More recently March 2026, Thaler vs Perlmutter, the Supreme Court ruled that AI may not be listed as an author for the purposes of copyright. That means that, should someone take your AI generated manuscript, put their name on it and sell it for their own profit, you will have no legal recourse to take the work (or subsequent profits) back, you’ll just have to suck it up, buttercup!
Ultimately the main selling point of your material is your passion for it! If you have striven and sweated over hundreds of thousands of words to tell the story that you feel the world needs to hear, then it is a passion that can be communicated, a torch that can be lit in readers that they can pass on in recommendations and (hopefully!) reviews. If the most you’ve done is push prompts into the plagiarism-machine, if you’ve not even read in enough detail to notice the AI prompt replies still within the manuscript, then are you really passionate about your art, or are you just looking to cash in as quickly as possible? Moreover, the audience response to cases where authors have been found to be using AI, and by extension the publishers’ responses, shows that, in actual fact, there is little appetite for AI aided or generated work. The audience overwhelmingly wants a connection with the author/creator, and the use of a machine is seen as a betrayal of that perceived connection.
One of the final failings of AI is that, by predating the wider samples of popular works it is inherently pandering to the lowest common denominator. Now don’t get me wrong, I enjoy a good Marvel Superhero Movie but, I enjoy one that makes me think! And, while we want our works to be accessible to the widest possible audience, sometimes we want to aspire to more complex concepts, higher philosophical heights. Sometimes we want to elevate the narrative and our readers by doing so. And, while the mainstream is mostly looking for the next X, insert derivative formulaic brain-candy here, we as authors can challenge readers while entertaining them at the same time. So please, exercise patience, hone your craft and, no matter what, leave the AI slop where it deserves to be, in the gutter.
David Cart-Writer 