Granta magazine asked Claude AI whether passages in a Commonwealth Short Story Prize entry were AI-generated and received opposite conclusions on rephrased prompts
Claude said it is not a detector and such checks are weak.
on the granta story. it’s clearly written by gpt. you can see all the motifs it loves and overuses like rain, weather, teeth, spine, memory. extreme overuse of figurative language and contrastive negation. it has the level of over-baking of probably GPT-5-thinking or 5.2-thinking
the story is … something ? I don’t think it has no value. the model develops an indo-Caribbean world register, man tries to murder his wife and chickens out. there’s some reasonable religious imagery where he combining three mythologies there with the names and whatnot
all of that is obviously overshadowed by the GPT prose style, and it’s hard for your eyes to not glaze over. there are various metaphors in there that boggle the mind. stuff like “the girl smiled like sunrise over a sink”.
what’s interesting is I went through the story and asked Claude Opus - a different model than the author model - and it seemed to find each and every one of the metaphors I hated brilliant. it finds a just so explanation for each of them when you press it
which makes you think, do these models have a shared internal vocabulary or compress various ideas in ways we don’t? the failures are quite interesting in that they reveal some different, and maybe bad, understanding of the human sensorium than a human has. why is pretraining knowledge compressed this way across all models? idk
#867Alexander Doria@DORIALEXANDERNew development: Claude just told me he officially disagrees with this framing, not what s/he meant.
Nabeel S. Qureshi@nabeelquHonestly hilarious that Granta's response to the whole affair is asking Claude whether the writing is AI, a method I would not consider reliable, given that: 1) LLMs tend to 'both sides' these types of questions 2) The way you frame the question to the LLM matters
@tszzl Our eyes glaze over because we've read enough similar text that it is low perplexity and taste for us, but if it was truly the first such story it would be kinda good, maybe?
roon@tszzlon the granta story. it’s clearly written by gpt. you can see all the motifs it loves and overuses like rain, weather, teeth, spine, memory. extreme overuse of figurative language and contrastive negation. it has the level of over-baking of probably GPT-5-thinking or 5.2-thinking the story is … something ? I don’t think it has no value. the model develops an indo-Caribbean world register, man tries to murder his wife and chickens out. there’s some reasonable religious imagery where he combining three mythologies there with the names and whatnot all of that is obviously overshadowed by the GPT prose style, and it’s hard for your eyes to not glaze over. there are various metaphors in there that boggle the mind. stuff like “the girl smiled like sunrise over a sink”. what’s interesting is I went through the story and asked Claude Opus - a different model than the author model - and it seemed to find each and every one of the metaphors I hated brilliant. it finds a just so explanation for each of them when you press it which makes you think, do these models have a shared internal vocabulary or compress various ideas in ways we don’t? the failures are quite interesting in that they reveal some different, and maybe bad, understanding of the human sensorium than a human has. why is pretraining knowledge compressed this way across all models? idk
This is truly embarrassing. I know not everyone is equally proficient or understand how AI works, but seriously, "I asked AI if it was AI" is worse than if they just left it up and stayed out of it.
If the judges had any artistic merit or ability they should give all the writers unlimited ChatGPT and then ask them to write a better story.
The problem isn't that they used AI, that's fine, it's that they used it badly and you're elevating lazy slop!
rohit@krishnanrohitExtraordinary response. As a famous award winning short story writer would say, this is not just constructing the funeral pyre, but actively lighting the match.
To demonstrate, here are two screenshots from Claude about the 'paragraph about the acre', which the Granta response cites as sounding human.
In #1, the model confidently argues that it's 'likely human'. In #2, with modified prompt, it confidently argues the paragraph is AI.
Nabeel S. Qureshi@nabeelquHonestly hilarious that Granta's response to the whole affair is asking Claude whether the writing is AI, a method I would not consider reliable, given that: 1) LLMs tend to 'both sides' these types of questions 2) The way you frame the question to the LLM matters
This sort of thing is obvious to people who have used Claude heavily and understand that models can be sycophantic, sensitive to framing effects, etc.
Nabeel S. Qureshi@nabeelquTo demonstrate, here are two screenshots from Claude about the 'paragraph about the acre', which the Granta response cites as sounding human. In #1, the model confidently argues that it's 'likely human'. In #2, with modified prompt, it confidently argues the paragraph is AI.
(response link: https://lunch.publishersmarketplace.com/wp-content/uploads/2026/05/Rausing-Statement.pdf)
Nabeel S. Qureshi@nabeelquThis sort of thing is obvious to people who have used Claude heavily and understand that models can be sycophantic, sensitive to framing effects, etc.