The AI Fiasco

[aworthyopponent]

It seems like many, many authors are struggling with getting stories rejected by the sites obviously too broad and near-useless AI filter.

At what point are we going to migrate to another site, and what sites would folks recommend for the exodus?

 

[Kelliezgirl]

We have no way of knowing if the filter is "too broad and near-useless".
Any system will have false positives. Without knowing the number of total rejections (and the number that are AI we don't have enough data to asses if the current system is or isn't effective.

That doesn't lessen the sting for people who are flagged incorrectly, but is the alternative to open the floodgates and let AI generated content take over the site?

 

[KeithD]

If my stories started getting rejected for that, I'd leave the site, yes. I've been posting here long enough and exhibiting technical dumbness on the discussion board long enough for the submissions editor to know I'm not using AI for my writing. If she lets an AI-check program have last word on acceptance of a couple of stories of mine (e.g., I've seen references to a story still being rejected after returning it with a pledge that AI wasn't used), I'll think she's lost control on selection. I'll leave at that point and only come back when I think she's regained control.

 

[alohadave]

It seems like many, many authors are struggling with getting stories rejected

We have no idea how many. The ones that come to the AH could be the only ones that are false positives.

There could be 10 times as many that never come to the forums to post about it.

There could be many more that are actual AI that are flagged for good reason that just go away.

We simply don't know the magnitude of the problem.

 

[KeithD]

But, when it pops up as multiple threads from multiple account names, chances are good it's a real issue.

 

[MrPixel]

There seems to be an avalanche of authors we have never heard from in AH that are reporting stories being kicked back for AI, some claiming "hundreds" of stories. I guess I'm too lazy to check them on that, and there's always the "alts" thing. As I posted elsewhere, I'm of two minds on this - is there too much protestation of innocence? Or is there actually a problem where Laurel is using an automated AI check with the threshold set too low?

Then there's the Grammarly thing, or even MSWord's grammar checker. Use Grammarly, go to jail, do not pass 'Go', do not collect $200.

Laurel certainly knows there's a problem. All we can do here in the peanut gallery is steel ourselves to the situation until something gives.

 

[OverconfidentSarcasm]

There seems to be an avalanche of authors we have never heard from in AH that are reporting stories being kicked back for AI, some claiming "hundreds" of stories.

Honestly? I kinda feel like it's been less than a dozen different authors I've seen complain in the forums during the last three months. But I do agree that the number of rejections is probably a lot higher than the number of complaints we see in the forums, as the number of authors active here is a lot lower than on the story side.

Maybe I'll just shoot Lazeez a message and ask whether there's been an influx of new users since the whole thing started here. After all, I think SOL would be the go-to alternative to Lit. It's not nearly as pretty as the new Lit design, and has nowhere near the reader base, but that would go for pretty much any other site.

 

[DDX84]

We have no idea how many. The ones that come to the AH could be the only ones that are false positives.

There could be 10 times as many that never come to the forums to post about it.

There could be many more that are actual AI that are flagged for good reason that just go away.

We simply don't know the magnitude of the problem.

I think the same. For all we know, the vast majority of cases are properly identified and rejected.

On the flip side, if there were hardly any cases, would Literotica need to implement new rules? No.

Besides, some of the authors have implicated themselves. Plenty of AH regulars were having discussions months ago about how much they loved using AI tools like ProWritingAid. It takes less than a minute to find them.

If I were Laurel, that alone would've been enough for me to put my foot down. If the best authors here are doing it, why wouldn't the beginners?

 

[djrip]

I've seen no evidence AI content is doing any harm to anyone here. We've all seen copious amounts of evidence that trying to reject it is doing lots of harm.

That's not to say AI content flooding the site couldn't do harm. We just haven't seen it do so.

If I were the site, I would modulate the rejection criteria to minimize harm in either direction. At this point that very clearly calls for reducing the number of AI rejections. Ie making it harder to call something AI and reject it. To address the false positives.

If at some point in the future AI getting through the filters begins to cause any sort of noticable problems, then adjust to make it more strict. Not as strict as it is now, but you know, respond to events as they actually happen. Instead of wildly rejecting vast swaths of what's being submitted out of a fear of AI when AI has not actually caused a single issue here, as opposed to efforts to stop AI which are causing tons of issues.

In my mind it's like guilt and innocence in a trial. I'd rather 10 criminals go free than see one innocent person imprisoned. Same thing for false AI positives.

 

[MrPixel]

Besides, some of the authors have implicated themselves. Plenty of AH regulars were having discussions months ago about how much they loved using AI tools like ProWritingAid. It takes less than a minute to find them.

Interesting. If I read those, I ignored it since I have zero interest in AI writing assists. I don't need some software engineer (I were one!) telling me how to write. Dr. McKimmey saw to that.

 

[AG31]

There seems to be an avalanche of authors we have never heard from in AH that are reporting stories being kicked back for AI, some claiming "hundreds" of stories. I guess I'm too lazy to check them on that, and there's always the "alts" thing.

I posted a relevant message in the AI Allegations thread (our attempt at a centralized, generic place to discuss this stuff.) Where I mention that I'd taken the trouble to look at earlier publications from some of the complainers. I invite you to pick up the discussion over there.

 

[BobbyBrandt]

As others had said, maybe the false positives aren't all that false. We have no way f knowing beyond what the writer claims.

I use the AI-assisted spell and grammar check within Word, and typically let Grammarly do a scan for something that Word missed, but these both relate to the basic typo and grammar faux pas any writer has to deal with. Nothing I have ever submitted here triggered any concerns about the use of AI.

 

[MrPixel]

I invite you to pick up the discussion over there.

No thanks. Once a thread gets much past 100 posts I usually bow out due to dilution from repetition and drift.

 

[djrip]

As others had said, maybe the false positives aren't all that false. We have no way f knowing beyond what the writer claims.

I use the AI-assisted spell and grammar check within Word, and typically let Grammarly do a scan for something that Word missed, but these both relate to the basic typo and grammar faux pas any writer has to deal with. Nothing I have ever submitted here triggered any concerns about the use of AI.

There is absolutely zero reason to believe Laurel has discovered the accurate AI detection method that no one else has figured out. No AI detector is much more accurate than a coin flip.

 

[Stimtheone]

I more often believe people are lying, as I have been using assisting AI for a while and haven't had a single rejection. Then again, I use it in moderation and only to correct my spelling rather than to generate, but the fact that what I consider to be a "normal" AI use has not had me rejected shows either these people are lying for using AI and not fessing up, lying intentionally to try and demonize AI without actually having submitted anything, or maybe they just want attention.

 

[djrip]

I more often believe people are lying, as I have been using assisting AI for a while and haven't had a single rejection. Then again, I use it in moderation and only to correct my spelling rather than to generate, but the fact that what I consider to be a "normal" AI use has not had me rejected shows either these people are lying for using AI and not fessing up, lying intentionally to try and demonize AI without actually having submitted anything, or maybe they just want attention.

You use AI and don't get caught, so the people claiming false positives are lying? No, that does not follow. What your data point actually does is add to the pile of evidence that AI detection does not work, since it failed to catch your use of 'assisting AI.'

 

[djrip]

I more often believe people are lying, as I have been using assisting AI for a while and haven't had a single rejection. Then again, I use it in moderation and only to correct my spelling rather than to generate, but the fact that what I consider to be a "normal" AI use has not had me rejected shows either these people are lying for using AI and not fessing up, lying intentionally to try and demonize AI without actually having submitted anything, or maybe they just want attention.

Wait wait, all you do is spell check? Then you aren't using AI at all. Either way, your experience in no way implies the many people getting false positives are lying.

 

[MillieDynamite]

We can post alternate sites to migrate to because that's against the rules. We'd have to share that in IMs or PMs or whatever communication acronym you would want to use. Other sites don't allow you to post links to sales sites, while some will let you link to your webpage. I don't plan on leaving or using AI, though I have thought about using AI to brainstorm story outlines. I've yet to try it, but if there is a use for AI where you still write your story, that'd be the one I'd be willing to try.

 

[MillieDynamite]

I went to answer someone and they removed their post.

 

[NoTalentHack]

Here’s a good example of why AI detection is nonsense (one, at least)
https://forum.literotica.com/threads/ai-rejection.1597700/post-98029539

TL;DR: I ran a section of my second story on the site, which was written only with Google Docs and no other tools, through 8 different AI detectors and got wildly differing results from “100% AI generated” to “completely human.” Even in the cases where they “detected” generated content, they couldn’t agree on which parts were AI written and which were human. All of them were human-generated, of course.

In the mega thread, I’ve given a bunch more examples. Sapling, DDX84’s weapon of choice, decided that almost every one of the top 10 most read stories on the site had significant amounts of AI content in them, even though the newest one was published in 2009. AI researchers say that the tools should neither be used as the sole determinant nor should they be used for anything important. They’re scarcely better than a coin flip, discriminate against non-English speakers, etc., etc. They’re about as useful as a witchhunter’s tests, and typically used for the same reasons.

Amazon basically gave up on AI detection, instead limiting authors to publishing no more than three books in a day. Medium has, too, relying on the honor system instead. There is no way Lit has found a better system than the ones they tried.

 

[samhasstories]

I have a hard time imagining that many people lying about stories getting rejected for AI. That's a pretty weird thing to do.

It is an interesting question though, as has been mentioned already: how many actual AI-generated stories are getting filtered out? Not having that information, it's hard to see if the blocks are actually a big problem, or just an unfortunate side effect to a necessary safeguard. I sympathize with the people getting their content blocked. But I also don't want to see this place inundated with AI-generated content.

 

[NoTalentHack]

As others had said, maybe the false positives aren't all that false. We have no way f knowing beyond what the writer claims.

I use the AI-assisted spell and grammar check within Word, and typically let Grammarly do a scan for something that Word missed, but these both relate to the basic typo and grammar faux pas any writer has to deal with. Nothing I have ever submitted here triggered any concerns about the use of AI.

This is my usage as well, but DDX84 and a few others claim it’s tantamount to cheating, so… Luckily, Laurel doesn’t seem to agree. However, the AI detection is still a terrible idea, and it’s clearly driving even established writers away.

 

[Stimtheone]

Wait wait, all you do is spell check? Then you aren't using AI at all. Either way, your experience in no way implies the many people getting false positives are lying.

Spell check and grammar check. And occasionally it corrects me on what words to use in a context since English is not my first language so it go "by ear".

Still, that is my opinion. I've seen a lot of pushback against AI, and while I understand their reasons I can't approve of the path. Though if people DID go 100% original and got rejected for AI it is a worthy cause, just feeling like there is a lot of heat around the subject.

 

[DDX84]

This is my usage as well, but DDX84 and a few others claim it’s tantamount to cheating, so… Luckily, Laurel doesn’t seem to agree.

Well, we both agree that Laurel's decision is final and I believe that she's made it in your case. I'm happy to accept that, as long as the rules apply to everyone. However, at the moment, that's not what's happening.

I haven't seen anyone else argue that the usage of AI in the form of ProWritingAid is tantamount to cheating. I'm happy to stand alone in that regard, but it's clear that my opinion is part of a small minority. Once again, if Laurel disagrees and thinks it's permissible, she should say so and include it in the publishing guidelines.

While I've reported stories and authors for various things over the years, your case was the only one which I reported for suspected AI use. However, as you've demonstrated in other threads, you've also been happy to report users if you felt the circumstances warranted it. Namely, the situation with Tilan, which resulted in others being banned that was based on your mistakes.

You've stated on multiple occasions that my report was also based on incorrect information. Yet, your reports were successful while mine were not. As has become clear, there are significant advantages to being Literotica's Most Influential Author. Congratulations on that, by the way.

However, it's neither fair nor accurate for you to describe your usage as being the same as @BobbyBrandt's.

I use ProWritingAid, tuned to the "Romance" setting with some custom choices in the advanced settings

I don't say this to attack you, but that's probably the most embarrassing admission I've ever seen a writer make.

View attachment 2305064

Do you understand why that's embarrassing? @onehitwanda's got the most impressive author's page I've ever seen. If I came here and discovered that the likes of her needed to do that, I'd die inside. I'd be mortified for her. All the great writers on the forum, insert the names of all your favourites. Do they need to lower themselves to doing that?

This is where you stand right now.

By your own admission, you rely on AI to conceive of changes to your final document, whether it's a single word, sentence, paragraph or page. Changes which are entirely conceived by the AI, not you as the writer. Changes which you allow the software to make on your behalf, after having first been tuned to the "Romance" setting. Then, you run a data visualisation to tabulate the words which you've used most often, to prop up an unsophisticated vocabulary, followed by more adjustments in pursuit of raising your "style score". For those unsure of what that last point means, the AI injects as much style and flair for you as it can, in alignment with the selected "Romance" theme.

At this precise moment, you lack the ability to edit and understand your own draft, while you also lack the ability to conceive of your own changes without a series of external prompts. Anyone lacking those elements from their skill set isn't a writer.

You don't have to be talented to contribute here. Every shit writer out there's welcome to come and have a go.

Just stand on your own two feet and write your stories yourself. It's really not a lot to ask.

 

[PennameWombat]

It seems like many, many authors are struggling with getting stories rejected by the sites obviously too broad and near-useless AI filter.

At what point are we going to migrate to another site, and what sites would folks recommend for the exodus?

On January 11, 2024, 158 - one hundred and fifty eight - new stories went live. It seems around 1200 total in the last seven days.

That's in keeping with the numbers when I've pulled stats for new stories for various reasons of curiosity.

IOW, whatever is behind the spate of AH complaints of AI rejections, they aren't having any sort of global impact in reducing the number of stories going live. As I have no information on how many stories have been submitted in the last couple of weeks, I can't state if there is any sort of rise or fall in rejections but the general average of stories going live hasn't changed.

Which means that "many" may or may not have any statistically significant meaning here in the total number of authors and submissions going through the grinder. It's not possible to know if the AH postings are 1%, 10% or 99% of the "AI" rejections.