Last updated May 30, 2024 | Author blog

You’re an author. You’ve worked on your manuscript for months. Years. A lifetime, even. But one day, you check the news, and a chill rushes through you. The Writer’s Guild of America has just announced that they’re going on indefinite strike. And one of the union’s main concerns? They want protections to guarantee that their work won’t be replaced by artificial intelligence. 

A month later, the screen actors guild also goes on strike — citing similar fears of AI encroachment. Suddenly, it’s all everybody and their dog is talking about: AI is a threat for creatives across all disciplines. Now, a year later, even though the strikes are over, a general aura of discomfort remains in the air amongst writers and artists alike. So the question remains: how concerned should you be about generative AI taking your job?

But first, how does generative AI even work?

Nowadays, all of our phones come with that fancy feature where it tries to predict and recommend the next word in your text message. However, I’m willing to bet that you never use it because it ends up generating nonsensical sentences like ‘Yesterday’s meeting is today, so we will have a meeting tomorrow before we go today.’ Genuinely, this is a real sentence my phone spat out when I tried to test this feature. If you overheard someone saying this across a busy office, maybe it would sound like a real sentence. But as soon as you listen carefully — and with intention — it’s immediately evident that it’s absolute gibberish.

That’s essentially the problem with how generative AI functions, and this is why it consistently seems to fail at writing good stories. It reads a vast amount of data from the internet, and then it attempts to map out patterns and predict what combination of text to spit back out. This normally works well enough if, for example, you want to find out what the most popular ice cream flavours are in Germany, or if you want to know the most common materials used to make working class shoes in 16th century England. AI can scan thousands of articles, studies, and polls, and then hand that information to you on a silver platter.

If that’s the case, why can’t AI write good fiction?

Because AI isn’t a real, human writer like you. It doesn’t have a personality, feelings, or strong opinions on whether skinny jeans are in or out. As a result, its interpretation of real-world experiences isn’t always processed correctly: just search up AI drawn images of human hands (or don’t, if you’re prone to nightmares), and you’ll see what I mean. An AI tasked to write a horror story will scan the internet for examples of horror stories, and it’ll simply copy what it sees. It may notice that ghosts and blood are frequently mentioned, and it’ll conclude that, rightly so, they’re quintessential to the genre. But it might also notice that these stories often describe characters sweating, and it could decide that sweating is, like blood and ghosts, something humans find horrifying.

I mean, mildly uncomfortable, sure. Smelly, definitely – but scary? And there’s the rub: AI can replicate what it sees in existing horror stories, but it’ll never understand the lived experience of feeling any of those fears itself. It doesn’t get that the smell of blood and the smell of sweat elicit very different responses. I might shudder at both, but absolutely not for the same reason.

Overall, the sentiment that AI written stories remain poorly constructed mimicry is largely echoed by writers across all industries. Authors, screenwriters, and video game writers have all experimented with letting AI generate their work. Rather unanimously, they agree that AI writing is, simply put, bad. In those rare cases when it isn’t incoherent, AI writing still relies on the lowest common denominator. Because it can only decide what should come next in its story by reading thousands of other stories, generative AI’s writing can only ever be, contrary to its name, reductive.

Join our email list for authors

Join our email list for content and opportunities to help you develop your writing, promote your books, and build your author business.

So we have absolutely nothing to fear, right?

Hold your horses! Unfortunately, like most things in life, the issue isn’t quite so simple. AI writing might not craft Michelin-starred meals, but it sure can pump out basic bread and butter. And bread and butter makes up a lot of our diets. Formulaic fiction is particularly threatened by AI’s encroachment because it involves creating stories based on familiar patterns.

The answer, then, is to ensure your writing stands out. Even if you write in a genre that leans towards the use of familiar tropes and formulas, make sure you put your own unique spin on them. Writers can also make use of the one thing generative AI definitely can’t manufacture – you! Build relationships with your readers by answering their emails, interacting with them on social media, participating in events or appearing on podcasts. Show them the person behind the words on the page, and it’ll make the experience of reading your stories more meaningful and unique.

Protection for writers may also develop in other forms. Legislation is being presented in most English-speaking countries to better regulate AI development and function, and many distributors and publishers are under increasing pressure to weed out AI-generated work in order to protect living, breathing authors.

For now, instead of worrying about AI replacing human authors, writers should instead see AI as the research assistant and responsive sparring partner that it can be. Its ability to comb the internet for small, technical details and improve research make it a valuable time-saving tool to add to the writer’s arsenal (just remember it’s not completely infallible!). In the same way that spell-checkers let us focus more on writing compelling narratives and spend less time searching up how to spell the word Ouija (seriously, those vowels?), AI will most likely end up as a writer’s assistant instead of the destructive, career-ending system that we all feared it could be.

Key Takeaway

In the end, it’s important for writers to remember that reading is not — and never has been — a passive activity. It takes work, effort, and imagination to absorb the words off the page, to digest them, and to create a mental image of the world being described. Someone reading a book is directly engaged in a tradition, spanning thousands of years, that links their imagination, through painstaking, thoughtful effort, with another.

Technology just can’t replicate this at the same level because it doesn’t understand what it means to be alive. AI-written stories break the sacred bond between author and reader; they break a connection where one person enjoys, knowing how much effort and dedication went into it, the imaginative work of someone else in another time and place. Even when a reader hates a book, that connection is still there — one that screams, ‘what were you thinking?’ Why would you do this to your beloved characters?’

Right now, AI is still in its infancy. And it’s on us to keep watch and make sure that developers and legislators parent it correctly. Like any child, it has the potential to grow into someone lovely, or it can become a difficult, fussy customer that ruins the day of everyone else in the restaurant. Let’s stay optimistic for now, and do our very best to raise it properly.

Dan Lin is the Coordinator for Fictive Pursuits, the company behind The Niche Reader, The History Quill, and Fabled Planet. Originally from Vancouver, Canada, Dan moved to London in 2022 to pursue a postgraduate degree in English Literature and a career working with others who are equally passionate about books. As a lover of the outdoors, you can typically find Dan on the weekends either rock climbing (when the English weather permits!) or exploring the green spaces around the city.

Join our email list for authors

Join our email list for content and opportunities to help you develop your writing, promote your books, and build your author business.