Kate's Newsletter logo

Kate's Newsletter

Archives
Subscribe
June 30, 2026

#25 AI and Fiction in the Wild

KP_author_branding_crop.png

Yet more AI writing and publishing scandals rumble on, with the Commonwealth Short Story Prize, and the Harper’s Bazaar short story prize the latest to come under fire. I hate to see writers disparaged and discredited; I also think this ‘us’ (those of us who don’t use AI tools) and ‘them’ (those of us who do) binary is not sustainable given how embedded these tools are in our smartphones and computers already. Who among us still writes with a typewriter? When is the last time you tried to hand in a manuscript written by hand?

In my last newsletter – ‘I Would Prefer Not to be Publicly Shamed’ – I was feeling optimistic about the potential for large language models (LLMs) to become another tool or set of tools for writers to use to counteract the hideous terror of the blank page. Now, once again, for a myriad of contradictory and confusing reasons, I’m not so sure.

spiky cactus in a shallow terracotta pot on a ledge with alien-looking pink flowers bursting up out of it
Metaphor!

One thing is clear and always has been: people need stories. You could say Keir Starmer’s demise as Prime Minister here in the UK came about at least partially because of his inability to tell a good story about his government and its achievements. We’ve heard a lot about the huge uptick in people using AI chatbots as therapists, but a recent academic paper startled me (though I am easily startled these days): it turns out a huge percentage of AI chatbot users – thirty-five percent in fact – are using these platforms to tell themselves stories.

‘AI Fiction in the Wild’ is an academic paper by Neel Gupta, Maria Antoniak, and Melanie Walsh, three American scholars from computer and information sciences. Here’s a quote from the abstract for the paper:

Some professional authors are beginning to use AI tools to help produce their fiction writing. Are readers using AI to generate fiction, too?

These scholars produced their research by looking at half a million English-language AI chatbot conversations. (Is ‘conversations’ the right word? Probably not). After conducting a deep analysis of these … transactions … they found that

more than one third of the conversations involve some form of fiction generation, including original stories, roleplay, fanfiction, and erotica.

Many of the users shared a similar profile, what the researchers call

infinite story demanders who repeatedly request and revise variations of the same or similar narratives over extended periods of time.

If you’ve ever taken care of small children you will recognise this plea: again. Again. Again.

The researchers used the data they collected to propose ‘two theoretical provocations’; firstly that AI technologies might lead a shift towards readers producing their own fictions that cater to their own tastes within a ‘closed conversational loop’.

Secondly, the researchers note that the AI chatbots

enable interactivity, play, and permutation in ways that are seemingly pleasurable for users, raising questions about where AI will fit into contemporary storytelling and entertainment ecosystems […] AI-generated fiction shares structural affinities with on-demand, personalized, and repetitive cultural forms.

Here the AI tool is like Scheherazad, whispering to the King – her own infinite story demander – night after night.

To reiterate what startled me: thirty-five percent of half a million interactions analysed by these scholars involved users working with the AI tools to craft stories, stories that these users wanted to be told, stories that catered to their bespoke narrative desires.

Your very own personal story machine.

Many people don’t interact with LLMs for ethical, moral and technical reasons. The tools impose limits on length both for inputs and outputs, so novel-length immersive long reads are not yet possible for the tools to generate. Many people value the community aspects of reading, whether that’s through belonging to a book club or a shared passion for fanfiction. And as all successful writers know, while good ideas are precious commodities, the ability to take that good idea and deliver a high quality novel based on it remains a challenging combination of graft, craft, luck and staying power.

So, will self-generated AI texts replace the novel anytime soon? Maybe. Maybe not. Who knows? Certainly not me.

🥳

In other news, I’ve started work on ‘Yudina’, my new opera commission, working with stellar composer Lubica Cekovska. My novel The Recent Dead is moving into the copy-edit phase with a Canadian publication date of next April. And I’ll be giving a keynote talk at Hypertext 2026.

Subscribe for free!
Don't miss what's next. Subscribe to Kate's Newsletter:
Older → #24 I Would Prefer Not to Be Publicly Shamed
Share this email:
Share on Facebook Share on Twitter Share on LinkedIn Share on Hacker News Share on Reddit Share via email
www.katepullinger.com
Powered by Buttondown, the easiest way to start and grow your newsletter.