AI Allegations Thread

[MillieDynamite]

AGH! Say it ain't so, No...

LOL that writer is me.

 

[ElectricBlue]

so it's like that judge who said back in the 1960s "I know obscenity when I see it"

Pretty much. It's not ideal, but you have to rely on your gut, because this stuff can't be reduced to a set of measurable rules.

Once people realise human creativity is an analogue art, not digital, they'll look at it differently. It's like you can hold a vinyl record in your hands, but you can't hold a bunch of streaming bits. Writing's like that, you've got to feel it.

 

[nice90sguy]

but you can't hold a bunch of streaming bits

True, that's why they invented pasta ladles.

 

[StillStunned]

True, that's why they invented pasta ladles.

And wank rags.

 

[ElectricBlue]

What I do with my stuff is my business. Those words are definitely mine, and I'll do whatever I please with them. If I want to throw in a few synonyms or mix up the order of some words, I'll do it; whether I do it solo or with a language tool. I'll freaking do it. My words, my business!

Unless I've misunderstood, you said you went with the suggestions the tool gave you, in the tweaked samples.

Or are you saying you re-wrote it, after seeing the AI critique, but with your vocabulary, not the AI suggestions?

I still reckon your first passes are the best, is all I'm really saying.

 

[nice90sguy]

And wank rags.

I still prefer to use a pasta ladle.

 

[Voboy]

But there are people who are not doing this that are getting denied and their previously published stories pulled en masse.

Yes, and I can well understand their stress. I've said a couple of times that Laurel is probably doing her best with this AI nonsense, but her best is no better than the "bests" of any number of institutions like colleges, the government, and other websites; NOBODY knows how to deal with this stuff properly. I hope she learns as she goes, and refines whatever she's using to detect this stuff.

As for those who use AI suggestions freely (@Dybbuk et al)? My take is this: if you claim you're using it to improve your writing, then I can see how that might be a valid justification in your eyes. But if that's your claim, you too should be learning as you go: you should be internalizing its inputs (since you seem to trust them), and you should be using it less and less with everything you write.

Can you honestly say this is happening? Or are you relying on it more and more as you "progress" in your writing?

 

[djrip]

but isn't everybody's gut going to say something different? Somebody posted parts of people's stories in this thread and said they were ai and then others read them here and said they weren't. And I totally get that it's up to the site owners but maybe if they let new people know their stories will be judged using ai detectors or by a human's gut feeling that would save a lot of this frustration. Right now the site just says don't use ai so people not using ai feel like they're following the rules when they submit a story. Then they get rejected And I don't get why one chapter would get pulled while the rest of the story is left up. That doesn't make sense if there's a reliable detector being used. I mean where's the consistency?

There is no reliable detector. It doesn't exist.

 

[alohadave]

Yes, and I can well understand their stress. I've said a couple of times that Laurel is probably doing her best with this AI nonsense, but her best is no better than the "bests" of any number of institutions like colleges, the government, and other websites; NOBODY knows how to deal with this stuff properly. I hope she learns as she goes, and refines whatever she's using to detect this stuff.

The silence is the most frustrating part. I know she's busy, and I'm not trying to dictate how she runs things, but when people are left to speculate in the dark, with random second hand accounts, they tend to assume the worst.

I also understand that the AH is a tiny proportion of the writers that submit stories, and we can be an echo chamber.

Even just knowing how many suspected AI stories she's dealing with might help to put things in perspective for us.

 

[nice90sguy]

Ok here's a challenge for any authors who HAVENT yet had a story rejected for possible use of AI. Write a story, without using any assistive tools (except spellcheck), that gets rejected as a false positive for AI. Then come back here and tell us how you managed to do it.

 

[DDX84]

I started using PWA June 5th. Since then, I've posted 14 stories and 1 essay, none of which encountered any issues in the queue, and two of which came in (probably, it's hard to be certain) fourth or fifth in contests. When I say that I take its suggestions "sometimes..."

Regardless, any additional changes are likely to be minor, and I've taken about all the suggestions from PWA I'm likely to take.

Like I said, I take, in a 10K document maybe 10-15 of these suggestions and swap them out for something else. But the rest? I let them ride. I do use some other tools, too; this is the bare minimum of what PWA can use.

In most other cases, I'm doing "better" than the average romance story (as determined by AC), or at least on par. But that "filler word" thing is killing me! So why don't I fix it?

If I drill down, this is the summary it gives of the filler words I use "too much."

Now, I did chop some out earlier where it was right. I had used "that" a few too many times where it could have been exempted when discussing action. The count was closer to 300 before that. But here are some examples that it wants me to remove:

It would read like a computer's idea of an English teacher's idea of people trying to navigate a difficult situation, at best. But my summary score would be better!

I wanted to thank you for taking the time to put that post together. I've found this thread to be really interesting for so many reasons, and I appreciate you clarifying parts of your workflow.

To be clear, I would never use tools like what you've described, even if the terms and conditions of the website allowed them.

When I sit down to write something, I want the finished product to have been my creation and for it to be a reflection of the sum of my skills and experience at that time. For me to test myself by creating something that different audiences would enjoy. If I've made small errors as part of those efforts, then it's up to me to invest the time to learn and improve.

Part of the skillset of being a writer is to read an early draft and detect the parts which need to be polished, tweaked or entirely rewritten. By subscribing to your raft of software solutions, you're asking the tools to identify those sections instead of doing it yourself. Each time you agree to let the software tools change your work, whether it's a single word, sentence or paragraph, you're learning nothing. Your skills as a writer aren't improving. It's lazy and goes against the very spirit of what it means to sit in front of a simple word processor and hone your craft, which takes time.

The specific examples you provided about your reliance on the tools to offer a quantitative analysis of which words you use too often, and the tools assigning you a score for style, are a terrible look. By relying on that data, you're failing to demonstrate that you possess the skills to make your own adjustments, or that you even understand them.

Tools like what you've described are used a lot in search marketing. Algorithms which tell copywriters and marketers how to write for search engines, then they quantify the effectiveness of the work by assigning a percentage score. That's not what this website is, or what it was ever envisaged to be.

However, beyond the fact as to whether those tools should be outlawed by Literotica or not, I just find it embarrassing why anyone would want to use them here.

The idea of spending hours upon hours drafting a story, or a novel, then letting a software tool make a variety of changes to it is embarrassing to think about. Letting the software tool seduce you into making those changes because your "style score" will jump from 55% to 75% is too embarrassing to even consider.

What this thread has made me appreciate, more than ever, are all the authors who know what it's like to sit in front of a humble word processor and put everything they've got into a new story. All those who do so after a long day at work, those who do so late at night and those who do so when they're not feeling at their best. Those who have to balance their writing with other responsibilities. Whether their writing is good, bad or anywhere inbetween.

Because they don't take shortcuts, and they don't cheat.

 

[NoTalentHack]

AGH! Say it ain't so, No...

LOL to be fair, it's before I've edited the back half. But from what I've seen, it's only going to change moderately. It seems to dislike my informal style of writing, which I'm okay with. Like, here's something it bitched about in the "showing versus telling" section:

View attachment 2294360

Now, I could probably tighten some of that up, and I might. Like, I could change "clamber over Aly to cross to the other side," but why? Who talks like that? It's dialogue between two normal people that aren't trying to give speeches, one of them (the narrator) talking about what he'd seen and how it made him feel, presenting it from his point of view, even though she'd been there, too. I might, maybe, change "feeling stupid even as I did so" to something like "mentally hitting myself" or something, but even that feels a little twee.

The previous part of the story was him sitting at a table by himself, reviewing in his head what had come before. That section is "fine" by the metrics that AutoCrit uses, as are later, more "actiony" ones during the foursome they eventually get into, even though I haven't done any real revising there yet; but I KNOW that part is weak.

This is a fine example of what I mean: neither this nor PWA "gets" what a writer is going for, no matter how "smart" their algorithms are. And trying to write "to" them instead of use them as a tool to evaluate is a danger. I'm not saying that would be true for everyone or for every story, but it definitely is here, and probably usually is for my stories.

Tools like this are developed to help hit metrics. Metrics aren't an inherently bad thing, but they can be. When I worked in tech support, we were evaluated for four things: time on call, number of times the customer called back, and number of parts sent plus... some other thing I can't even remember now. But each of those was a reaction to the previous one.

Time on call?

"Whoops, guess I hung up."

So they added in a metric for number of times the customer called back.

"Fuck, I don't know. Send him all six things it could be."

Okay, add in number of parts sent as a metric.

"Shit. Entire new computers only count as one part, right?"

I've always hated the slogan, "you can't manage what you can't measure." Because as soon as you can measure it? Its usefulness as a management tool drops like a rock. People start targeting the metric instead of the end goal, because they're being rewarded for hitting or failing to hit the metric. And all something like AC does--at least for people who aren't trying to write commercially--is tell them "this is the metric, aim for this," when they should be thinking "this is my goal, and here are some metrics. Do they align?"

I'll put it another way: if AC had come out in the 1900s, it would have wanted flowery, ostentatious prose with way too much... stuff. It would have told Hemingway he sucked. The tool would have been geared towards making things that were commercially successful IN THAT DAY, because that's all metrics can be tuned towards. It's the same thing people say about AI, that is completely accurate: it's not going to create anything new. It's only going to refine what exists.

 

[nice90sguy]

I read one of @NoTalentHack 's stories. I enjoyed it. But there again, I am an AI myself, so clearly I'm biased.

 

[NoTalentHack]

I wanted to thank you for taking the time to put that post together. I've found this thread to be really interesting for so many reasons, and I appreciate you clarifying parts of your workflow.

To be clear, I would never use tools like what you've described, even if the terms and conditions of the website allowed them.

When I sit down to write something, I want the finished product to have been my creation and for it to be a reflection of the sum of my skills and experience at that time. For me to test myself by creating something that different audiences would enjoy. If I've made small errors as part of those efforts, then it's up to me to invest the time to learn and improve.

Part of the skillset of being a writer is to read an early draft and detect the parts which need to be polished, tweaked or entirely rewritten. By subscribing to your raft of software solutions, you're asking the tools to identify those sections instead of doing it yourself. Each time you agree to let the software tools change your work, whether it's a single word, sentence or paragraph, you're learning nothing. Your skills as a writer aren't improving. It's lazy and goes against the very spirit of what it means to sit in front of a simple word processor and hone your craft, which takes time.

The specific examples you provided about your reliance on the tools to offer a quantitative analysis of which words you use too often, and the tools assigning you a score for style, are a terrible look. By relying on that data, you're failing to demonstrate that you possess the skills to make your own adjustments, or that you even understand them.

Tools like what you've described are used a lot in search marketing. Algorithms which tell copywriters and marketers how to write for search engines, then they quantify the effectiveness of the work by assigning a percentage score. That's not what this website is, or what it was ever envisaged to be.

However, beyond the fact as to whether those tools should be outlawed by Literotica or not, I just find it embarrassing why anyone would want to use them here.

The idea of spending hours upon hours drafting a story, or a novel, then letting a software tool make a variety of changes to it is embarrassing to think about. Letting the software tool seduce you into making those changes because your "style score" will jump from 55% to 75% is too embarrassing to even consider.

What this thread has made me appreciate, more than ever, are all the authors who know what it's like to sit in front of a humble word processor and put everything they've got into a new story. All those who do so after a long day at work, those who do so late at night and those who do so when they're not feeling at their best. Those who have to balance their writing with other responsibilities. Whether their writing is good, bad or anywhere inbetween.

Because they don't take shortcuts, and they don't cheat.

Okay, most of this made me LOL. "Simple word processor." I don't know how old you are, but I predate the PC word processors. I remember writers back then bitching HARD about how word processors were cheating because they let you revise so quickly. Before that, the bias was against typewriters, because they let you "write faster than you could think." (Not joking, I remember seeing that in a very old magazine article).

The way I use the tools are no different than having an unpaid, fairly inept editor. It underlines stuff and says, "this is bad wrong no good stuff," and I tell it to go fuck itself, because who's paying who here? Or I listen to it, make the tweaks I feel like (and I am making them, I'm not asking it to rewrite stuff for me), and go on down the road.

And if you'll look up above to my response, I actually agree about writing to the metric. Metrics are meant to be tools, at least when it comes to artistic stuff, not hard and fast rules. I can and do ignore them if I think they're wrong, just like I do editors that I disagree with.

 

[Voboy]

Part of the skillset of being a writer is to read an early draft and detect the parts which need to be polished, tweaked or entirely rewritten. By subscribing to your raft of software solutions, you're asking the tools to identify those sections instead of doing it yourself. Each time you agree to let the software tools change your work, whether it's a single word, sentence or paragraph, you're learning nothing. Your skills as a writer aren't improving. It's lazy and goes against the very spirit of what it means to sit in front of a simple word processor and hone your craft, which takes time.

You're pretty much reading my mind. If you're an AI, I'll be highly abashed.

The idea of spending hours upon hours drafting a story, or a novel, then letting a software tool make a variety of changes to it is embarrassing to think about. Letting the software tool seduce you into making those changes because your "style score" will jump from 55% to 75% is too embarrassing to even consider.

"Embarrassing" is how I look at it, too. I'd feel a real sense of shame if I tried to pass off AI as my own work. I suppose I can appreciate that not everyone feels that way, but I think it's a firm line, and you either "get it" or you don't.

I appreciate that @NoTalentHack posted that stuff. It's important for all of us to know what we're talking about and see examples of it, otherwise we're all just old men yelling at clouds.

 

[nice90sguy]

Have you ever actually checked what "looking like AI" means? It's pretty simple. Just make sure your style differs from that. And if it doesn't, change it. You really don't need to be too precious about "Your style". Eveybody changes styles over time anyway.

 

[DDX84]

You're pretty much reading my mind. If you're an AI, I'll be highly abashed.

"Embarrassing" is how I look at it, too. I'd feel a real sense of shame if I tried to pass of AI as my own work. I suppose I can appreciate that not everyone feels that way, but I think it's a firm line, and you either "get it" or you don't.

I appreciate that @NoTalentHack posted that stuff. It's important for all of us to know what we're talking about and see examples of it, otherwise we're all just old men yelling at clouds.

Yeah, I think you've captured the essence of the issue there. There are some who get it and those who feel that it's no big deal.

Okay, most of this made me LOL. "Simple word processor." I don't know how old you are, but I predate the PC word processors. I remember writers back then bitching HARD about how word processors were cheating because they let you revise so quickly. Before that, the bias was against typewriters, because they let you "write faster than you could think." (Not joking, I remember seeing that in a very old magazine article).

The way I use the tools are no different than having an unpaid, fairly inept editor. It underlines stuff and says, "this is bad wrong no good stuff," and I tell it to go fuck itself, because who's paying who here? Or I listen to it, make the tweaks I feel like (and I am making them, I'm not asking it to rewrite stuff for me), and go on down the road.

And if you'll look up above to my response, I actually agree about writing to the metric. Metrics are meant to be tools, at least when it comes to artistic stuff, not hard and fast rules. I can and do ignore them if I think they're wrong, just like I do editors that I disagree with.

With respect, what you're arguing here is all over the place. Word processors are not the same thing as AI tools which generate content. The speed with which someone types has got nothing to do with the creation of content by AI tools.

Going back up to what Voboy alluded to, you've just proven his point. There are those that get it and others who don't even have so much as a fundamental understanding of the issue.

And you don't tell your tools to "go fuck themselves", because you clearly stated earlier that you've made a large number of adjustments to stories which you've written since 05/06/2023. Adjustments which were conceived and implemented by software tools and published, by you, as entirely your own work.

You're talking about "artistic stuff", here's a genuine question for anyone reading this. What's artistic about using software tools and AI to radically change a manuscript at the touch of a button? Whether it's changing words, sentences, paragraphs or entire pages.

On that note, how do you all feel about entering into a contest here on Literotica with someone who's just freely admitted to doing exactly that, while creating stories on an industrial basis?

 

[NoTalentHack]

And you don't tell your tools to "go fuck themselves", because you clearly stated earlier that you've made a large number of adjustments to stories which you've written since 05/06/2023. Adjustments which were conceived and implemented by software tools and published, by you, as entirely your own work.

You're talking about "artistic stuff", here's a genuine question for anyone reading this. What's artistic about using software tools and AI to radically change a manuscript at the touch of a button? Whether it's changing entire words, sentences, paragraphs or pages.

On that note, how do you all feel about entering into a contest here on Literotica with someone who's just freely admitted to doing exactly that, while creating stories on an industrial basis?

Apparently, reading comprehension isn't one of your strong suits. I literally said above that I've NEVER let the software rewrite my stuff. What I've done is paid attention to where it said, "I don't like this, and here's why," weighted whether I thought it made sense in the context of the work, then made a change myself, if I felt it was correct.

To repeat myself, I have never, not once, done this:

View attachment 2294384

View attachment 2294385

Instead, I looked at that and decided that I liked "very obviously" and stuck with it. In a case where I wouldn't, I would have tried to come up with a different adjective and, failing that, gone looking at thesaurus.com to find one. Or is that cheating, too?

Here's the problem: you don't know what the fuck you're talking about, but you're trying to enforce rules around it. That is always, always, always a recipe for disaster, no matter the topic.

I've probably thought more about ML and its repercussions legally, ethically, practically, and artistically than 90% of the folks on these forums. I've spent my entire adult life in software development, automating tasks. I have literally destroyed people's jobs without meaning to and laid awake at night thinking about that.

I've watched IDE (integrated development environments) go from dumb text editors to sophisticated tools that can evaluate what a dev is writing and predict what they'll want next, knowing that each improvement was coming closer to putting me (or more likely a junior developer) out of a job.

I also knew that it was only a matter of time before English-language text would come under fire.

I understand, at least on a basic level, the process of what's going on in Stable Diffusion, and what we can guess at in ChatGPT, since it's a closed algorithm. And I also know that all this discussion we're having, all the handwringing and "is this art" and the various debates on how one should feel proud or shamed is just spitting into the wind.

I write because I like to write. I wrote for fun before I used any automated tools. My use of tools has streamlined my process, but also slowed it down, but it hasn't fundamentally changed what I've written. Even once the tools are good enough to fool any detector (which they will be eventually), I'll write because I like to write, and I still won't use tools to replace my own writing. But that's my line in the sand, and it's all going to be on the honor system.

I just don't act like it's some great moral stand because I do. Because I know where this is headed, and no one's going to give a shit by the time the next generation is writing their stories. Or, not writing them, if you choose to believe that.

 

[XerXesXu]

Have you ever actually checked what "looking like AI" means? It's pretty simple. Just make sure your style differs from that. And if it doesn't, change it. You really don't need to be too precious about "Your style". Eveybody changes styles over time anyway.

No one knows what 'looking like AI' means. How and What AI learns is inaccessible, other than it learns a root-mean-square of HI text. HI text will differ from that to a greater or lesser extent - think bell curve. Those HI writers who fall closer to the median decile will contain more of these unknowns than those who fall further from the mean. They're unlucky, nothing more than that.

 

[TheArsonist]

"Embarrassing" is how I look at it, too. I'd feel a real sense of shame if I tried to pass of AI as my own work. I suppose I can appreciate that not everyone feels that way, but I think it's a firm line, and you either "get it" or you don't.

From the example @NoTalentHack posted, I wouldn't consider that to be "passing off AI as his own work" at all. It's his writing, the AI is only providing analysis and suggestions. Even if he were to accept a suggested rewrite, IMO that's not much different from accepting spellcheck's suggested fix of a typo, or clicking a word to bring up a built-in thesaurus and changing it to a synonym, or using autocomplete when typing on a phone: the choice is made by the writer, the software just streamlines the task by proposing alternatives.

Sure, if someone blindly applies AI suggestions en masse without review or critical consideration, then they are no longer actively engaged in the writing—much as if they blindly accept edits by an editor: they're abdicating the role of author to a co-author. I thought the interesting part of his post was how it showed that that's not how these tools are usefully employed.

 

[TheArsonist]

No one knows what 'looking like AI' means. How and What AI learns is inaccessible

There's also the complication that not all AI-generated text is the same. This is particularly relevant to erotica. ChatGPT and many other prominent tools are specifically designed to not produce NSFW outputs, so anything of that sort people manage to get them to produce is by definition abnormal. I suspect that this makes AI-detectors even more unreliable.

 

[MillieDynamite]

What author or genre are you having your work compared to? You don't have to lose everything in red. But you might want a few synonyms to sprinkle in. Using the words felt, fell, and felling is considered to be telling and not showing. Your repetition level wasn't that bad. I don't exclude names or anything, so my repetition is always high.

LOL to be fair, it's before I've edited the back half. But from what I've seen, it's only going to change moderately. It seems to dislike my informal style of writing, which I'm okay with. Like, here's something it bitched about in the "showing versus telling" section:

View attachment 2294360

Now, I could probably tighten some of that up, and I might. Like, I could change "clamber over Aly to cross to the other side," but why? Who talks like that? It's dialogue between two normal people that aren't trying to give speeches, one of them (the narrator) talking about what he'd seen and how it made him feel, presenting it from his point of view, even though she'd been there, too. I might, maybe, change "feeling stupid even as I did so" to something like "mentally hitting myself" or something, but even that feels a little twee.

The previous part of the story was him sitting at a table by himself, reviewing in his head what had come before. That section is "fine" by the metrics that AutoCrit uses, as are later, more "actiony" ones during the foursome they eventually get into, even though I haven't done any real revising there yet; but I KNOW that part is weak.

This is a fine example of what I mean: neither this nor PWA "gets" what a writer is going for, no matter how "smart" their algorithms are. And trying to write "to" them instead of use them as a tool to evaluate is a danger. I'm not saying that would be true for everyone or for every story, but it definitely is here, and probably usually is for my stories.

Tools like this are developed to help hit metrics. Metrics aren't an inherently bad thing, but they can be. When I worked in tech support, we were evaluated for four things: time on call, number of times the customer called back, and number of parts sent plus... some other thing I can't even remember now. But each of those was a reaction to the previous one.

Time on call?

"Whoops, guess I hung up."

So they added in a metric for number of times the customer called back.

"Fuck, I don't know. Send him all six things it could be."

Okay, add in number of parts sent as a metric.

"Shit. Entire new computers only count as one part, right?"

I've always hated the slogan, "you can't manage what you can't measure." Because as soon as you can measure it? Its usefulness as a management tool drops like a rock. People start targeting the metric instead of the end goal, because they're being rewarded for hitting or failing to hit the metric. And all something like AC does--at least for people who aren't trying to write commercially--is tell them "this is the metric, aim for this," when they should be thinking "this is my goal, and here are some metrics. Do they align?"

I'll put it another way: if AC had come out in the 1900s, it would have wanted flowery, ostentatious prose with way too much... stuff. It would have told Hemingway he sucked. The tool would have been geared towards making things that were commercially successful IN THAT DAY, because that's all metrics can be tuned towards. It's the same thing people say about AI, that is completely accurate: it's not going to create anything new. It's only going to refine what exists.

 

[MillieDynamite]

Also, @NoTalentHack, get is always considered a bad word along with that, big, bad, and, I guess, wolf. No kidding, not the last word. As in, "That big bad wolf just ate granny."

 

[XerXesXu]

there are a lot of unlucky bastards in this thread

You'd expect more people to be nearer the median than further from it. Visualise the area under the curve.

 

[MillieDynamite]

@NoTalentHack, this is AutoCrit's review of a recent 9,000-word story of mine.

View attachment 2294396

Please don't think I'm saying I write better than you. I don't. However, I do use the software available so I can make my writing better. And yes, the goal I have is that those purchasing my stories will like what they read and return for more. I write for me, and there are many who can't stand what I write about. But hopefully, I'm building some followers at sales sites that will buy my stuff and, when I fall short, or even flat on my face, give me the benefit of the doubt that it was just a bad day.