People are using AI but not disclosing it. How do you detect it ?

[aquinas]

You want some grass fed fanfic? Just $49.99 for 1000 words. Certified Humane!

Click to shrink...

Honestly, might get to the point where we're paying people to live-stream writing Asuka X Hikari fics on a Selectric, or the insane net-filters out of Stephenson's Anathem.

 

[BronzePlaceWriter]

And everyone is gonna lie about it.

In some games the only winning move is not to play at all.

Click to shrink...

True, but less people than you might think.

Saying that you're AI free when you're not is an active lie. It requires you to intentionally take action so as to decieve.

Simply not declaring yourself Ai is a passive lie. You don't actually have to do anything - it's much easier to justify to yourself than going out and intentionally misleading people. Of course, there are people who still would, quite a lot of them, but don't underestimate it as a filter. Plus, if someone actively advertises themselves as AI free and later its found out they're not? The backlash can be a lot greater because again, they actively choose to lie. People don't like that, especially readers.

 

[xihute]

I don't care if they are ridiculed or that their reputation is at rock bottom, I won't go out to defend them or attack them personally or support witch hunts or do anything that may affect anyone directly.

But of course I have opinions on the people using AI and the state of AI generated work.

Click to shrink...

Okay.

It's just that from the general tone it sounded more like "I don't care because I agree they deserve it" than simple "I don't care".

Just explaining my take. No smoke.

You want some grass fed fanfic? Just $49.99 for 1000 words. Certified Humane!

Click to shrink...

And everyone is gonna lie about it.

Click to shrink...

Honestly, might get to the point where we're paying people to live-stream writing Asuka X Hikari fics on a Selectric, or the insane net-filters out of Stephenson's Anathem.

Click to shrink...

At the start of ChatGPT craze I've read a short story "Writer looks for readers" or similar title.

The premise was that with advent of AI written fiction at quality of high-level writers, accessible immediately, about anything a reader wants, demand for human-written stories plummeted.

And so, the job of reader for hire was created, but even this some people faked with AI, and in the end we have scene face to face between hired reader and amateurish writer who wants someone read they work - and only face to face she can be certain.

So, the story missed the mark, but maybe not completely. By hiring an author to write a story face to face you can be sure that it is "Certified Humane".

The poor plebs will be having AI-generated slop, and rich snobs will have personal writers.

 

[The Radiance's Follower]

Okay.

It's just that from the general tone it sounded more like "I don't care because I agree they deserve it" than simple "I don't care".

Just explaining my take. No smoke.

Click to shrink...

Nah, it's fair. I thought about it for a few minutes. You are right. I don't care about them because I think any ridicule they get from videos on Youtube or articles written about them is deserved. I don't waste my time thinking or interacting with them. I want to have close to no interaction with any 'art' they make.

It is the reason why I started the thread in the first place. I want to find the most expeditious way to recognize if a story is AI and drop it and move on with my life. There are thousands of stories made each and every day to read from that are not AI. And even more stories made even before AI.

 

[Sailor Autochthonia]

I'm not involved in many large communities, but what I've seen from smaller ones is that uh, generally the people posting AI content are just people who like AI. Not a cabal of sinister hooded figures trying to get one over on the poor innocent general public.

And like most people who are people, they tend to try to avoid getting harassed about, let's say, literally anything. Which generally means not sticking out from whatever "normal" is. I'd bet that if we had a dedicated subforum for AI generated content, even if it wasn't the exclusive, enforced place for AI content, most people would post their AI generated content there, because it's just a nice cozy treefort for people with a shared interest, rather than, like, "Hey, hang out with everyone else, but wear this big neon sign that X% of people here vocally hate!"

Even if rule 1 is enforced just as hard against people being dicks about AI in general CW as it would be in this hypothetical AI CW, the vibe's different. "My feelings don't care about your facts" deal. Plus, similar to what BronzePlaceWriter said, anyone being a dick in AI CW would be actively going out of their way to be a dick, making it easier to dismiss them as some sad pathetic loser with nothing better to do than go into the AI section and bitch that they found AI. Again, even though that's just as true about a story in general CW tagged as AI... my feelings don't care about your facts.

Then again, if we look at the number of SFW stories in NSFW (take a shot) and the reason for, we might just end up with that situation again, idunno, I think the mods already said they wouldn't make another subforum, but I might be hallucinating, on account of how I'm actually— uh, ignore that last part.

 

[Spudr]

If that is a scam, then video editing, photo editing and others done to make people look prettier than what they are is a scam.

Click to shrink...

You really love making false equivalences, don't you?

 

[xihute]

Sidenote: it's irritating how many threads about AI there's already, and that they all converge. It looks like the same conversation is in multiple places and it gets disjointed.


I have pretty important question: how many people care about AI in writing? Do we have some inkling?

The labeled AI-writings get decent number of likes, despite having all hellmarks of AI, and there's an author that disclosed use of AI (going by the story I've read he clearly put more effort into that, I would say 70/30 at least) that gets >1k.

 

[Lightbringer]

You really love making false equivalences, don't you?

Click to shrink...

For someone who keeps spouting 'ah buh this is different' merely to support a illogical position?
Sure, call it 'false equivalence'.
They did argue in the past that photography would kill painters, might as well see that they had a point, sure?

 

[Lightbringer]

Anyway.

I don't care about AI for a reason. I have readed a fic, it was a Worm/Dune crossover, and up to a point I suspected it was AI generated.

Problem? I can't tell if it is truly AI generated since nobody seems to think that and so on. So I had two avenues of inquiry.

1. Assume the guy's actual writing is not that high.
2. Assume he is lazy.
3. Assume he is a good writer.

The issue with the second? It still requires effort so lazy is out of the window.

We go back to the first option, but that comes on another issue for me.
Assuming that his writing is in truth not as good as we see, then I would not read it. Which gives the reason why he used AI in first place. Which is irrelevant because then I would have not readed it because it is AI.

But if he didn't use AI, then his writing is good, which would mean I would read it.

Then I noticed I cared too much if the writing is good to ponder if a human did it or not. Who cares? Most humans write awful badly written stories, if that means less awful writing forms and spelling, oh well, I will not complain about it.

Nonetheless said fic I abandoned it because Paul is not like that.

 

[Spudr]

For someone who keeps spouting 'ah buh this is different' merely to support a illogical position?
Sure, call it 'false equivalence'.
They did argue in the past that photography would kill painters, might as well see that they had a point, sure?

Click to shrink...

I suppose you extend that to false advertising in general, then? Or claiming that a food product does not contain nuts when it does, in fact, contain nuts?

Or is AI just a special case because... reasons?

 

[Lightbringer]

I suppose you extend that to false advertising in general, then? Or claiming that a food product does not contain nuts when it does, in fact, contain nuts?

Or is AI just a special case because... reasons?

Click to shrink...

No? If anything I wanna see if they would extend to those as well.
It is the logical end of that inquiry.
But we know the crowd will just make some excuse for it and AI is the 'special case' where that inquiry fits...for some reason.

 

[doll]

Or claiming that a food product does not contain nuts when it does, in fact, contain nuts?

Or is AI just a special case because... reasons?

Click to shrink...

The final product doesn't contain AI, AI is just used in the manufacturing.
A more accurate analogy would be freerange vs. cage eggs, but even that's wrong because the actual product does differ depending on how you keep livestock.
Closer I think would be when clothing products put the country of origin on the tag. If you care about that sort of thing (e.g. you buy local products preferentially) maybe the actual product isn't the question but rather who gets paid for what in the process of buying it. Of course this is a flawed analogy at a more basic level because commercial fanwork is itself not the norm in the first place, but it's not too far off spiritually - it's a question of who made it and how rather than what it is.

 

[Spudr]

No? If anything I wanna see if they would extend to those as well.
It is the logical end of that inquiry.
But we know the crowd will just make some excuse for it and AI is the 'special case' where that inquiry fits...for some reason.

Click to shrink...

As the saying goes, perfect is the enemy of good. We're not talking about changing laws, we're talking about setting the policy of this forum.

 

[LurkingInTheDeceit]

it's due to the engagement that sites like these benefit from having AI software crank out chapters on the daily.

Click to shrink...

We aren't really putting any effort towards growth, and we don't have advertising, so we aren't giving AI use free rein to increase engagement.

I don't like AI written stories either, so I'd also like it if I could filter them all out. It's just not possible to implement that, and any attempt would likely be a poison worse than the disease.

I have. My solution is taking a clear moderation stance against undisclosed AI usage. Doing so would lower unlabeled AI usage overnight.

To expand on that, I would introduce a checkbox when posting new stories similar to the checkbox for the necroposting warning. Explaining that undisclosed AI usage is against the rules.

I want to emphasize that I understand your gripes with this solution, but even if it's just this and the moderators take no proactive enforcement (unless presented with undeniable evidence) undisclosed AI usage would lower by a huge margin. You only need to see the effect of the necroposting warning has on you to see the correlation.

Click to shrink...

'Don't make rules you can't enforce', as the saying goes. But say we did display that policy in various places, I agree it likely would reduce the amount of untagged AI use, toothless as it would be.

It would also give license to people who think they've discovered untagged AI use to flood the report system (which, fine, we can just auto-reject them), or try to 'righteously' argue with an author/other readers that the work is definitely AI because of X/Y/Z, and should be tagged as such. Making it official policy (even if unenforced) would just encourage the witch hunts we don't want even more, and that's not a small problem as is, when the official policy is solely 'we don't care, stop harassing the author'.