70 comments

[ 1.5 ms ] story [ 37.5 ms ] thread
Very bleak. I think like most people, I would prefer for stories not to be written by AI. I value human creativity, and believe that an LLM cannot truly produce art.

This article suggests that when people are lied to that stories were not written by AI (when they actually were), they prefer them.

"Art is in the eye of the beholder" - it is me, who enjoys art and how it affects me and works with my emotions that is affected by it. Who or what the artist is, should be irrelevant, unless the artist imposes some monetary value I want to profit from at some point.
I don’t think the artist should be irrelevant - the art is the expression of the artist’s emotions. Knowing about the artist unlocks a lot of depth in the art, it’s an attempt to share and express their perspective through creative means.

Not to say how you decide to consume art isn’t valid (there is no proper definition after all). However, you’re doing a disservice to a lot of artists by dismissing their role in it all.

It does - but does this not potentially suggest that an LLM can produce "art"?
To me "art" is the communication of something from the artist, through arrangement of reality, to the person experiencing the art.
For practical questions, I always prefer Claude's answer to a human's. Non-narcissistic and to the point. For entertainment, I'm less impatient, although unsurprisingly, most people view reading as homework and are happier to blow through a book written at a YA level. Brandon Sanderson a'int the king for nothing.
It should be noted that these stories were 1000 word stories, with each one taking 5 minutes to read on average.
If it's not already the case, it likely will soon be that people prefer short main motifs for songs generated by AI over humans.

It could still be quite some time before they're writing entire songs and books, though.

No one buys 1000 word stories, or listens to 12 second songs on Spotify, so for the time being it's not really any cause for concern.

If Facebook has demonstrated anything it's that people have preferred AI memes for a while already.

People "prefer" lower quality art all the time - look at what's big on the radio and in theaters, what's popular on TV, etc.

Why should I be concerned if AI is making trash art instead of humans?

I love Britney Spears when I was 9. Is the world really worse off if I'm listening to an AI version or Toxic or the Britney Original?

I can see why people would be scared if AI was pumping out Purple Rains and Superstitions on the regular, if it was writing a Secret History every hour, but it's not, nowhere even close.

This just seems like people being grumpy that people prefer Marvel movies over what they think are good movies.

No one really knows how this will play out, but it's still uncomfortable to think about robots figuring out how to reliably make things that look good, sound good and are full of meaning, would people still have any motivation to create? Would anything be considered good if everything is "good"? Of course things could turn out fine, I just wouldn't minimize the significance of what's happening.
This is already the struggle for 99% of people that would like to be artists (or athletes, or a number of things)...
Yes, people will invariably have the motivation and drive to create.

We didn't stop playing chess when Deep Blue beat Kasparov. We don't stop swimming because boats are faster. We don't stop running, even though we can't outrun a car.

Humans do these things because they are gratifying to do. It is satisfying to train for and then run a marathon. It's fun (and rewarding!) to challenge yourself to pull the image out of your mind and place it on a canvas, an invitation to the inside of mind.

I don't care if The Beatles are considered a better band; my bandmates and I simply like making music together. It's fun.

>It should be noted that these stories were 1000 word stories

A huge limitation of the study IMO.

Most people who read are not reading 1000 word short stories. If they could replicate this story using novella length stories (25-40k words), I'd be curious to see the results. I would wager that a much larger % of people would prefer long form content from humans.

Rare exception of Betteridge's law of headlines
They are not very representative. I absolutely hate the subjects and plots and writing all over and would have voted randomly.
Honestly much more curious to know about the human written stories were here! Feels at least pretty important to the study. With the AI, the specifics feel less important, because the relative input is going to be so small, so the researchers can probably just do it themselves. But how do you choose which works to stand in as pieces of human writing in general? How did they make sure the subjects weren't aware of the authors they chose?
Yes! Reading all samples really drives home the study's conclusion that “people generally prefer predictability, because difficult material requires more brain power.” I could blow through the AI generated versions as I do Brandon Sanderson or JK Rowling because my brain was saying "yada yada yada, yeah I get it, how many sentences do you need to say the same thing."
I cannot judge fairly given that the stories are labeled, but the story about the koi is dreadful. It is about a supposedly mysterious subject—the machine at least insists it's mysterious—but it is so dull, so pedestrian, so cut-and-dry. "My mommy dying makes me sad but the peaceful koi pond helps me find peace". Blech.

Perhaps it only seems so banal because the competition is so engigmatic.

And what does it mean for something to smell of "sterile detachment"?

What was the quality? AI generally will have better grammar and such vs what most people write. So I could believe you prefer the AI content in this example because it is better quality, but this tells us nothing about AI vs human written when someone takes the time to edit their rough drafts.

(Note that hacker news discussions are not conductive to spending days crafting a good response - that takes days - so everything here is a rough draft)

Of course AI is hyped at Villanova:

https://www.villanova.edu/university/ai.html

"Villanova embraces Artificial Intelligence (AI) with a commitment to thoughtful, ethical use rooted in our Augustinian Catholic mission and values. We aim to balance innovation with our humanist tradition and intellectual heritage, preparing students to lead responsibly in a changing world.

Explore resources designed to help you learn, teach and work with AI at Villanova."

Funny. It's the same result as that famous painting labeled as AI-generated and people found many faults with it. It's just showing that people are biased against AI itself; the quality of the content doesn't really matter.
The real question to be debated is how do you deal with a new kid on the block, who is better than you at [whatever].

Denial seems to be the general human default, it's easy and free, but at best it can only delay the inevitable.

Accept and adapt is probably the most logical approach, but that requires a lot of effort and risk taking, and still will likely leave you net worse off.

I suppose hostility and violence is another human favorite, but that almost never has gone over well.

I don't think reading isolated stories holds much value. The value of the emotion it elicits is much lower than knowing the person who wrote it or in what historical or social background they did so.

Reading stories written by humans is one of the ways we can keep ourselves grounded to other humans: be it through fantasy, lyrics, or autobiographies. I can only describe reading AI stories as emotional indulgence.

> The value of the emotion it elicits is much lower than knowing the person who wrote it or in what historical or social background they did so.

Then the value of many ancient texts must be close to zero, because we often know nothing about the people who wrote them. In many cases even our guesses for where and when they were written differ by multiple centuries or thousands of kilometers.

No that makes them more interesting for the same reasons OP gave. You try to build a mental model of this mystery person from a mysterious time, where ever little detail could be a clue about them, their society/culture, or just trivial.

With AI writing the answer is that everything is randomly averaged and thus trivial.

> You try to build a mental model of this mystery person from a mysterious time

I’ve read many ancient texts with great delight, and not once have I tried to “build a mental model” of their authors. It’s a dimension of literary analysis that has never interested me, even for authors (such as Kafka) whose personal lives are generally considered inseparable from their works.

The work stands on its own for me… or doesn’t, in which case learning about the circumstances of whoever wrote it won’t save it in my eyes.

Every story told, is folded into the corpus of all human stories, all variations of the heroes journey, all variations of the hammon-cycle. AI by being trained on the corpus, is average - which makes it better than 80 % of mankind.

But we are tired of many stories. We see "new" movies until we are around 20 - then the repetition sets in- we get a refined taste and long for the exotic (-30) until finally, we search for outliers, for anything new, desperate under the sun.

AI can not provide that. But also humans can not provide that. Every bard, every writer, every poet, every band rises every morning to compete against the beatles. Most of them loose and thus abandon the craft to win bread elsewhere.

> But also humans can not provide that. Every bard, every writer, every poet, every band rises every morning to compete against the beatles. Most of them loose and thus abandon the craft to win bread elsewhere.

Most, but not all. For every 99 who give up and go and do something else, one transcends.

AI cannot create. Humans can. Most do not, but the possibility is always there that they could.

Not something novel, usually - but i can reproduce recombination results with intuition, which is kind of close enough. Still it lacks the striking direction of an artist/great mind and averages out.
but most of what humans create isn’t novel, esp in saturated areas like art. so even LLMs can “just do that” they’re very capable at outputting worthwhile materisl
"most of" being the key words.
Intuition is just detecting and following patterns, which AI is already very good at.
Now show me an AI coming up with General Relativity, or Integral Calculus, or Chopin's wrong note Etude.

AI builds on what comes before. Humans have an uncanny trait for creating things that are truly new.

Can't exactly show since those things have already been discovered of course. But there is already a good list of AI discoveries out there in various fields, and it's still early days.

Humans also build on what comes before. Many of the "truly new" things that humans have created happened by accident, was discovered by brute force/experimentation, or are remixes of things observed in nature and/or already done by others. There is literally nothing a human can invent that a sufficiently capable AI can't also invent as the techniques are the same.

If you seriously dig into literature or film you'll find that many of the stories don't fit into the hero's journey at all--or only fit if you make the "journey" so generic as to be practically tautological.

As for the claim about people's lives. People who are serious about art almost invariably keep enjoying the "basic" classics; many musicians keep listening to Bach all their lives. Faulkner claimed to read Don Quixote over again every year. It is possible to find novelty in the well-worn.

In short: creating stories requires mental effort, and humans don't want mental effort, so use AI.
I’m in both a lot of art based communities and programming based communities (like HN). The one blind spot many HN readers have is that the quality of something is one facet of an artwork.

Humans make minute decisions in their art whether to tell a story or represent something. When someone generates something, at some point you as the viewer/reader could assume that arbitrary choices were made in places. Those arbitrary decisions just aren’t interesting. I’ve quoted this before, but Kane Parsons highlights exactly this thing:

> If I see an element of the environment has been, you know they used generative fill or whatever to change something about the scene, even if--. It just shuts the part of my brain that wants to know more about that world and wants to look for details, 'cause I would assume if they're willing to make an arbitrary choice there, they can make an arbitrary choice with literally anything.

It’s just less interesting to our minds, I think because there’s less to ponder and explore. It can be a masterpiece in quality but when I know something in that piece was just made up by a machine, why would I care? I can’t wonder why the artist would include it. I can’t think about what it’s trying to tell me because it’s likely not telling me anything.

I think it’s a blind spot to a lot of people not just programmers. If you don’t have experience understanding and not just viewing you’ll almost always prefer the shinier object. I don’t really think this is particularly new even generative AI has just ramped up the speed that shiny objects can be produced. Most people aren't trained to do the understanding part so most media has always been about maximizing the shininess.
In other words, most people have terrible taste.
That's why those "we prefer human writing" polls (such as the one cited in [0]: whooping 98% readers say they prefer human writing to AI!) are just noise.

To quote myself:

> People are really, really good at lying to themselves, let alone to an online poll. They'll tell you that they prefer imperfect or even bad writing as long as it's not AI slop, just like how they'll tell you they like healthier food, they prioritize personality instead of look for potential dates, how they use LLM "only as a spellchecker", and how they use tiktok for educational videos. As long as there is no stake, people will just say what make they feel better.

[0]: https://bcantrill.dtrace.org/2026/09/05/the-revolt-of-the-re...

> That's why those "we prefer human writing" polls (such as the one cited in [0]: whooping 98% readers say they prefer human writing to AI!) are just noise.

Why would they not?

> They'll tell you that they prefer imperfect or even bad writing as long as it's not AI slop,

But I do, and it's very clear that the online communities I frequent prefer people write in their own voice, even with very bad grammar and spelling, rather than resort to chatgpt for their posts and comments — this comes up all the time

> just like how they'll tell you they like healthier food

But lots of us do? Once you are over 30 and you control your own fridge and cook your own food, why would you not?

> they prioritize personality instead of look for potential dates

But (in a fair context for comparison) people actually do, especially women. As long as someone is presentable and shows a bit of self-care and self-respect, for the most part looks rank at least on a par with personality, if not below it.

> how they use LLM "only as a spellchecker"

I don't use it at all for writing; I do know people are weak on this point but I also know that people understand it fosters distrust

> how they use tiktok for educational videos

Huge chunks of people don't use it at all.

Most of what you are saying here sounds like you are projecting low mood and your own personal prejudices and insecurities. Try to be less misanthropic.

Yeah why not?

Srsly.

Nonethless, a lot of people enjoy stupid movies, others enjoy smart movies etc.

If you are not like me and you don't care about the time you spend on a 2h movie because you just want to watch something, you will have a complete different experience than me. I watched a 1000 movies and more, I want/I need something 'new' or unique i haven't seen a 1000 times.

Its the same with the AI slop videos; I found the first 3 'cats with jobs ai' videos funny as a category, i don't want to watch thou 'accidents' videos if they are fake.

I'm sorry but all this proves is that a lot of published stories aren't very good. The three human stories attached don't have fantastic plot and read a little poorly.

All three of the human written stories read like they were made for a reviewer/publisher rather than a reader.

I don't think this is a good study or a good representation of good writing. I think many short stories have the issue of force feeding so much information and tricks that the 3 human stories read more gimmicky than the chatgpt ones.

This is a bit like asking "do people prefer vegetarian meals if we don't tell them there's actually meat in it?"

People care about the production process even if they can't tell you're lying to them about it.

It's quite likely that many survey participants were AI bots, and hence preferring AI generated content. And certaintly, that of the 1600+ people whose 15min they rented, few were paying any meaningful levels of attention to earn their 3 USD.
"We assume creative writing requires uniquely human qualities, such as emotional understanding and lived experience, which leads people to underestimate AI's capabilities."

It's not about the relative "creativity" of it. Art is fundamentally about communication. It's a social activity, whether you're reading a book or listening to music or looking at a painting. Even a mid-average author is going to make something more compelling, on a human level, than extraordinary prose generated by a machine.

AI-generated creative artifacts are as interesting as the fractals in a snowflake. Interesting, sometimes beautiful, but void of human meaning.

Is art what is generated, or is art what is in the eye of the beholder? Do we find artistry in nature?

What does it say about us that the sum average of us disgusts us on a visceral artistic level, is that the machine's fault for reflecting us accurately?

Art is a social activity whereby a human uses their creativity to produce an artifact that expresses some aspect of their experience, using non-standard forms of communication. It's a way of signaling to others "do you see what I see? do you feel what I feel?" The beauty of art exists on two layers: the aesthetic quality of the work, and the sense of shared experience among those who consume the art.

We find aesthetic beauty in nature; we do not find artistry.

The “sum average” of people is a bad caricature, not “reflecting us accurately”.
I'd like to see the results sliced up for an age range of roughly 35-70. I would heavily question the taste and judgment of those outside that range. Or at the very least note that people's personalities will change significantly when entering/exiting that range.
These are ~1000-word stories (prompts at [1]), squarely in the sweet spot where planning logistics are minimal, context windows comfortably fit everything, and the appropriate level of abstraction omits any details which could glaringly reveal weak world-model knowledge.

SOTA LLMs from 2 years ago would saturate this study, and while modern LLMs are amazing and I suspect they would do quite well on a similar but more challenging study that forces them out of said comfort zone, this isn't that.

[1] https://www.cambridge.org/core/journals/judgment-and-decisio...

No, I have 0 interest in AI stories or music or art. I like all of these things because it's a view into the mind of the artist and I want to connect with another human and appreciate what they made and why.
LLM writing is optimized, at a great expense, for that specific parameter. They literally have human raters fine-tuning "favorability" - general appeal - as a part of post-training. So yes, the conclusion here doesn't surprise me at all. By the way, this is also true of HN! You think you're better than that, but at this point, on any given day, we have around 10 LLM-generated stories on the front page (and another 10 are usually AI-related press releases or op-eds).

Anyway, human writing isn't like that. You have a style that some people will like, some people will find "meh", and some people will hate.

A better question is, do people want to read AI text all day? And the answer to that question is, I think, very different. Most people don't. Which is why so many parties are really determined not to tell you.