[object Object]
How Human Does AI Fiction Actually Read? We Built a Way to Measure It
Introducing Fiction Slop Index!
The AIStoryHub blog
[object Object]
Introducing Fiction Slop Index!
Most writing communities split into two camps. The purists, where using a model gets you treated like you turned in a counterfeit. And the volume crowd, where the words are just what stands between you and the upload button. I belong in neither, so I built a room for that.
We built AIStoryHub on a stubborn idea: that you can write with AI and still sound like yourself. This August we are putting money behind it. The AIStoryHub Short Story
Story Rules and Presets can now reach directly into your story's own codex at the moment a chapter generates.
Twenty-five real AI clichés, pulled straight from the 758-entry corpus behind our Slop Checker. Click a square when you spot one in the wild, a LinkedIn post, a press release, the fourth paragraph of a ChatGPT answer.
Nine words keep showing up right after a comma in AI prose, restating what the sentence already said. Here's why it's so hard to self-edit out.
A 742-entry corpus of measured AI writing tells, and why the same five clichés show up in generated fiction no matter who's prompting it, or which model wrote it.
No more subscriptions. Let's build the community of writers!
We gave four AI models identical writing prompts and scored every output with a blind judge. The models all wrote publishable prose. The judge scored itself 8.8. GPT-5.4 invented a dead mother we never mentioned.
There's always a gap between what you can see and what the page actually says. The only way to close it is to hand the work to someone who isn't inside your head — but getting a draft to the right readers, without exposing it to everyone, has always been harder than it should be. Today that changes.
Chapter One of the novel The Midnight Lounge showcasing what's possible on the platform.
Win an Amazon Gift Card!