![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
asakiyumeAccording to Neil Clarke of Clarkesworld (as quoted in this Guardian article), there are get-rich-quick schemers out there who are encouraging people to submit AI-generated stories to high-paying, highly regarded venues. Clarkesworld has seen an increase from an average of 10 to over 500. As a consequence, Clarkesworld has closed submissions.
Another venue said it would only accept submissions from known authors.
That’s a terrible blow for up-and-coming writers and ultimately for the whole ecosystem. How to solve it?
First, I want to clarify the difference between the problem as it exists now and the ultimate problem. Judging from the fact that Clarkesworld was able to recognize and reject 500 stories as AI-generated, the problem right now isn’t that AI-written stories are indistinguishable from human-written ones; right now it’s a problem of spam. It’s a problem of a deluge of trash submissions making it untenable for zine teams to sort through to find the genuine ones.
Ultimately, as AI-generated stories get better, we’ll have the problem of distinguishing them from human-produced ones—if we decide that's a problem—and the solutions will be different, but I have some ideas for right now.
Idea 1: a cool-off period. Writers submit their names only. They are contacted a month later and invited at that point to submit their story. This ought to deter most spam.
Idea 2 a change in directionality. What if instead of authors submitting to publishers, publishers went looking for authors? This is already what’s had to happen to increase submissions from marginalized, lesser-heard-from demographics: publishers have actively sought them out. It’s distressing for writers to have to sit around like flowers in a garden waiting to be picked, but it’s a possibility.
Idea 3: writing circles. Essentially groups of writers who choose to come together to write in a certain style or about certain topics or just because they get along. They share writing with one another, talk about and share stories they’ve read as well. They would share some writing publicly (for free), so that there would be a public record of the circle’s existence and the sort of work its members produced. Then once every [time period], circles would make recommendations to zines of works to consider for publication. In other words, writers themselves would be doing first-level slush management, and zines could judge the types of stories they’d likely be getting from the circles by the work posted publicly.
These ideas have drawbacks, I realize, but maybe with refinement one or several of them could work?
Another venue said it would only accept submissions from known authors.
That’s a terrible blow for up-and-coming writers and ultimately for the whole ecosystem. How to solve it?
First, I want to clarify the difference between the problem as it exists now and the ultimate problem. Judging from the fact that Clarkesworld was able to recognize and reject 500 stories as AI-generated, the problem right now isn’t that AI-written stories are indistinguishable from human-written ones; right now it’s a problem of spam. It’s a problem of a deluge of trash submissions making it untenable for zine teams to sort through to find the genuine ones.
Ultimately, as AI-generated stories get better, we’ll have the problem of distinguishing them from human-produced ones—if we decide that's a problem—and the solutions will be different, but I have some ideas for right now.
Idea 1: a cool-off period. Writers submit their names only. They are contacted a month later and invited at that point to submit their story. This ought to deter most spam.
Idea 2 a change in directionality. What if instead of authors submitting to publishers, publishers went looking for authors? This is already what’s had to happen to increase submissions from marginalized, lesser-heard-from demographics: publishers have actively sought them out. It’s distressing for writers to have to sit around like flowers in a garden waiting to be picked, but it’s a possibility.
Idea 3: writing circles. Essentially groups of writers who choose to come together to write in a certain style or about certain topics or just because they get along. They share writing with one another, talk about and share stories they’ve read as well. They would share some writing publicly (for free), so that there would be a public record of the circle’s existence and the sort of work its members produced. Then once every [time period], circles would make recommendations to zines of works to consider for publication. In other words, writers themselves would be doing first-level slush management, and zines could judge the types of stories they’d likely be getting from the circles by the work posted publicly.
These ideas have drawbacks, I realize, but maybe with refinement one or several of them could work?
no subject
Date: 2023-02-23 02:01 am (UTC)Which is what's so rough about the situation. A few bad faith actors with some time and technology on their hands, making things harder for so very many of good faith actors.
But anyway, yes, very true about idea 1! (And for all my little self-pity party moment up there, I do really appreciate that you're looking for productive ways to address the issue, whereas I've just been hoping that the people who run magazines figure it out. Not optimally helpful!) There could also be green M&Ms clauses of sorts, if you know what I mean -- something to include in your submission file to show that you read the directions, that changes venue to venue or even month to month within a venue. I do really like the idea of little tests like that that shouldn't take writers much effort beyond reading the submissions page, but would weed out the spammers who can't be bothered to do that.
no subject
Date: 2023-02-23 05:05 am (UTC)