A surge of AI-written novels submitted to literary awards — leaving judges scrambling to keep up — became a hot topic on 5ch’s Hardware/Industry (Console Wars) board. Some in the thread pushed back, arguing that AI prose, easy to read without any friction, just doesn’t suit literature meant to be read between the lines. Others countered that, used well, AI can produce writing that’s clear and easy to follow. The discussion eventually grew into a debate over just how much human involvement is needed for something to still count as “creation.”
The Age of AI Novels Flooding Literary Awards: The Judges’ Struggle and the Truth Behind “Readable Writing”
Why are AI novels shaking up the awards selection process so much? Behind it lies more than just a rise in submission numbers — there’s a psychological mechanism at work in which humans end up genuinely preferring AI-written prose.
Source: news.yahoo.co.jp / Read the original article here
if only it were even that good
Brain-dead affiliate spammers (アフィカス, ad-farming blog copy-pasters) can only copy-paste anyway
Guess AI needs to start the threads too now
That's basically the point, right?
AI doesn't actually write from scratch as much as people think
you need a solid plot and detailed instructions or it won't give you the writing you intended
used well though, it can turn out clear, easy-to-read prose
I read literature specifically because I want to spend time and brainpower on it, the last thing I want is plain, easy-to-digest prose
if I just wanted simple content consumption I'd watch Netflix instead
If the content's actually interesting, I'll happily read between the lines all day though
Writers who grow up on AI will end up writing in an AI-like style
language is always changing anyway
Fixing it is overwhelmingly faster [than writing from scratch]
with grunt work, you're still the one in the driver's seat
with the latter, you're letting AI replace even your brain
it's supposed to be your own means of expression — outsourcing that to someone else defeats the whole point
there's an old story about a noble who had a painter make a picture for him and then bragged about it as if he'd painted it himself, and got mocked for it. this is exactly that
the quality of the output actually changes depending on the instructions and prompts you feed the AI, so the personality of someone who uses it properly really does come through
that noble made literally the same excuse you're making now lol
something like 'I'm the one who told him to paint the seaside, or the birds, or whatever, so this painting being good is thanks to me too' lol
but having AI take it from 0 to 1, then the creator develops it from 1 to 10
or the creator does 0 to 1 themselves and has AI take it from 1 to 10
I think we're just at a point where denying that is pointless
even for literary awards, I bet the people who do the entire 0-to-10 process completely by themselves are going to become the minority as time goes on
nah, in the end novels written entirely by humans still sell better, so that's not happening
lol that's dumb
the correct way to use AI is:
have AI do a full 1-to-10 draft
then, with real intent, the human either does it themselves or re-prompts the AI
to condense that 10 down to a 5
then again, human or AI, re-prompt
to condense that 5 down to a 3
and using that 3 as a base, the human
rewrites it back up to about a 6
only then do you actually get quality control
unfortunately most people
just stop at 1→10→5
meaning they never actually revise
and honestly, if a human writes it entirely themselves
they usually can't even manage a full 1-to-10 in the first place
which used to be exactly the filter that weeded out most people
the judges are AI
the readers are AI
wait a second…
did we just build a perpetual motion machine?
motion to self-destruct: seconded
it'll happen
but even if some amateur like you uses AI to write a novel, you're not winning any award with it
well sure, if your level is 'change the sentence endings and suddenly it's a different character,' then maybe that's true for you guys
yeah it'll happen
because it's all about connections anyway
the ghostwriter just becomes AI instead
in other words, it's a trash industry
not worth paying into
even before that, it was already nothing but ad-stuffed corporate sites
these days when I'm looking something up, I've stopped even bothering with sites updated in just the last few years
too many garbage sites that are just AI copy-paste jobs 🙀
AI just says things that sound plausible, after all
it's not grounded in real experience, so it rings hollow
sorry for not saying anything sooner
guess it's about that time again — the annual server-reset season
ha, 'AI-onion' — that's a good one (a pun combining AI with onion, riffing on the 'simulated world' joke above)
Background and Key Points
Japan’s major literary prizes — the Akutagawa and Naoki Prizes for established writers, and feeder awards like the Bungakukai Newcomer’s Prize or the Shincho Prize for New Writers — all rely on a first pass by human readers who cull thousands of unsolicited manuscripts down to a shortlist before a committee of established authors makes the final call. That screening layer is unpaid or low-paid and was never built for a world where a submission can be generated in minutes rather than drafted over months, which is the practical crunch the article’s headline refers to. Notably, one prize, the Hoshi Shinichi Award, has explicitly permitted AI-assisted entries since 2013 and even shortlisted a Nagoya University-linked AI-assisted story in 2016 — most literary awards, by contrast, have no such rule and are having to improvise.
The thread’s actual fault line wasn’t “AI good or bad” but how much revision counts as authorship: one camp (post 72/81) accepted AI drafting as legitimate provided a human meaningfully reworks the output (their “0-to-1, 1-to-10” framing), while the opposing camp (post 42, invoking the nobleman-and-painter fable) argued that outsourcing the sentence-level expression at all forfeits the claim to have “created” anything, regardless of how good the prompting was.
What the thread never touches is the “psychological mechanism” the article’s own intro promises to explain: processing fluency — the well-documented cognitive bias where text that’s easier to read gets rated as higher-quality or more trustworthy independent of its actual content. That’s a plausible reason AI prose keeps clearing anonymized screening rounds, but no poster raises it.
*This article is compiled and summarized from the 5ch (Hardware/Industry (Console Wars)) thread “[Sad News] Literary Awards Are Getting Flooded With AI Novels lmaooo.”
