The phrasing “bias against AI” seems to beg the question here. The article takes it for granted that people are wrong to say they’re more interested in stories written by people than by AI, because they can’t tell the difference if they’re misled.

Compare with a hypothetical study saying: people say they prefer true inspirational personal stories to fake inspirational personal stories. But if you lie to them, they think the fake ones are just as good!

Obviously, this would not prove that they are “wrong” or “biased”. The whole point of stories written by people is that a _person_ wrote it, based on their actual human thoughts and experiences.

> The whole point of stories written by people is that a _person_ wrote it, based on their actual human thoughts and experiences.

I thought the whole point of stories was that they were entertaining or had some pertinent message.

I imagine at the time of the printing press, someone argued "The whole point of stories is that a person wrote it"

I had this experience while watching the last season of Black Mirror. I asked GPT for an episode and the outline came out much more interesting than anything on this season [0].

I may be in the minority here, or at least am willing to admit my preference, but I really don't care where my media comes from. I don't believe I have some special relationship with artists. I want people to have jobs and live a fulfilled life, but I don't feel any loyalty to something like listening to music made by humans or basing my consumer decisions based off things like royalty reimbursement rates. If I like it, and there's an incentive for its creation, I'm fine by that. People are still free to create art. I write pretty regularly but I don't expect to be able to do that for a profession, and that's fine.

[0] https://mleverything.substack.com/p/we-should-just-let-gpt-w...

The title is misleading, this study doesn't even attempt to actually answer preference of human vs AI written stories. It didn't even involve any human written stories!

All this study shows is that people rate an AI generate story worse when they know it's AI and say they'd pay less for it. Who knows what the relative rating or "payment" for a actual human written story would have been!

A cynical but imo largely correct take on this phenomenon: "People prefer A.I. art because people prefer bad art" https://maxread.substack.com/p/people-prefer-ai-art-because-...

(TL;DR - AI art is by design an attempt to reach the lowest common denominator, which makes it easy to digest, and a lot of people like stuff that's easy to digest.)

I question whether AI can create engaging new ideas. Isn't everything it outputs already based on existing data?

That's true but we have to ask do humans actually create engaging, new ideas? When you look at all the stories that have been created aren't they just slight variations on the same themes that have been around for, as near as we can tell, all time? When we look back at our historical writings it's the same themes at their core over and over it's just a different way to tell the story. Considering that we can look back 10,000 years in history and see the same common themes being retold it is quite likely that prior to that we were still telling the same stories.

We can find different ways to tell them which can be entertaining and it can feel new but when you examine it it's not really new.

Whereas this comment, seen fifteen thousand plus times on this very website, is evidence of ... your human creativity?

The amount of work done that is truly new is laughably small. A large percentage of creative output is convergent (combining two existing things) or lateral (bringing an idea from one space to another).

The problem of originality has been of intense debate since at least the romantic era.

There are people who have argued that human creativity is nothing more than outputs based on existing data (experience).

Title needs to be changed. It completely misrepresents this research. There was no comparison between human written and AI written stories.