Best current guidance on Literotica and AI

[MillieDynamite]

For the AI determination software to work correctly for you, you'll have to buy a subscription. There are five or six metrics that are used in the full blown packages. Well beyond what we use to determine what's what for our own work. (speaking of me, not you). I'm not buying software to tell me I'm not a robot writing a story. But I'll not complain about them wanting human-written stories and using whatever they choose to determine it. We, if we don't want to go to the trouble, don't have to publish here. They don't need to cater to our whims, proclivities, or writing forms.

 

[MillieDynamite]

Let me say this about that, Grammarly helps with my spelling, points out my transpositions (think dyslexic), and grammar. However, sometimes I use bad grammar intentionally. I take or rejection suggestions based on what I'm doing with what I write.

@StillSunned in response to

If I may ask, why do people use Grammarly in the first place? Because they're uncertain about their grammar?

 

[Bamagan]

I'm a little surprised that a professional editor is advising not to use tools to double check the text.

I was equally surprised to hear from a doctor that it was pointless to use WebMD, but at least I didn't have seven types of cancer. And I got a free lollipop!

 

[alohadave]

I was equally surprised to hear from a doctor that it was pointless to use WebMD, but at least I didn't have seven types of cancer. And I got a free lollipop!

Hold up. WebMD isn't accurate? Maybe I shouldn't have have my spleen taken out after all.

 

[StillStunned]

I'm a little surprised that a professional editor is advising not to use tools to double check the text.

Like any professional, I'll advice against any injudicious reliance on tools over basic knowledge.

I myself use a spellchecker. I use Read Aloud for the final proofreading. In my professional work, I'll almost always have a colleague check my work.

There's nothing wrong with using a tool to check your work. But people here are saying that they rely on Grammarly to point out things they didn't know were wrong. Fair enough, if that's how you want to learn. But make sure that you're learning the right lessons - not business English for fiction, for a start - and that you need to be learning the lesson in the first place.

If you can speak English, you can write it. That's one of the beauties of the language: it can be fantastic at every level of proficiency. Don't overcomplicate your writing beyond what's comfortable for you just because you think that's what "real writers" do, or worse because a computer program tells you to.

 

[SimonDoom]

If I may ask, why do people use Grammarly in the first place? Because they're uncertain about their grammar?

It's very simple. It's not because I'm uncertain of my grammar knowledge, because I know that my grammar knowledge is very good. It spots things I might not. I might know grammar, but that doesn't prevent my fingers from doing weird things as I type. Tools like Grammarly, or Word's editing tools, will catch things and show them to me. I can then decide whether and how to respond to what's caught.

I think the crucial step is not automatically accepting what these tools tell you to do. Use your judgment.

Also, this may sound funny, but I think it may be true and useful: make an active effort to write like a human. If you do, I bet you'll avoid these AI false positives most of the time.

 

[ElectricBlue]

Also, this may sound funny, but I think it may be true and useful: make an active effort to write like a human. If you do, I bet you'll avoid these AI false positives most of the time.

Agree this. From reading these multiple threads, I can't help but think many writers have absorbed too much of the "grammarly style", which might be great for business writing but rubbish for fiction. It's a misapplication of what might be a useful tool - if you're writing bids and proposals.

 

[RejectReality]

Also, this may sound funny, but I think it may be true and useful: make an active effort to write like a human. If you do, I bet you'll avoid these AI false positives most of the time.

Make an effort to write like a human? Is that really the best way to phrase that?

I get where you're coming from, but that just comes off as elitist snobbery of the highest order to me, and I'm not even dealing with false AI rejections.

 

[SimonDoom]

Make an effort to write like a human? Is that really the best way to phrase that?

I get where you're coming from, but that just comes off as elitist snobbery of the highest order to me, and I'm not even dealing with false AI rejections.

I don't know why that would seem snobbish. I mean by that to write with quirks and idiosyncrasies and changes that would make one's writing seem less programmed. Mix up sentence structure and word choice. Write like one is talking rather than like an essay. I don't know exactly how AI works but I'm guessing it's looking for patterns, so as one writes one can try to avoid patterns that would make one's writing look too smooth or too canned.

I realize that this is a bigger problem than I originally had thought, but it definitely seems to be the case that some people are having a much bigger problem with it than others, and that means that there are things one can do as an author to minimize the problem. It's not snobbish or elitist or dismissive to point that out that this is a problem that to some significant degree probably is within one's control. I have never had this problem with any of my submissions. I also do not ever use AI in my writing.

 

[overwatcher]

Make an effort to write like a human? Is that really the best way to phrase that?

I get where you're coming from, but that just comes off as elitist snobbery of the highest order to me, and I'm not even dealing with false AI rejections.

I don't know why that would seem snobbish. I mean by that to write with quirks and idiosyncrasies and changes that would make one's writing seem less programmed. Mix up sentence structure and word choice. Write like one is talking rather than like an essay. I don't know exactly how AI works but I'm guessing it's looking for patterns, so as one writes one can try to avoid patterns that would make one's writing look too smooth or too canned.

I realize that this is a bigger problem than I originally had thought, but it definitely seems to be the case that some people are having a much bigger problem with it than others, and that means that there are things one can do as an author to minimize the problem. It's not snobbish or elitist or dismissive to point that out that this is a problem that to some significant degree probably is within one's control. I have never had this problem with any of my submissions. I also do not ever use AI in my writing.

__________________
Ever kinky and curious,

SimonDoom

Yep. I agree with Simon, adjust sentence structure and the words formed, from the normal and carry it over into the dialogue as well. All of that helped my last two stories.

I'm from down under so sentence structure has to an extent always been different from the norm, even in dialogue. Aussies and Kiwi's speak faster so sentences are longer don't embody punctuation even for emphasis and we complete sentences with words like 'but' to show we're finished and use 'Look' far to often to start sentences. So words used, phrases, colloquialisms and cultural references are embedded in my stories.

Ah and swearing - we do a lot of that, moreso than North America and Canada (I've worked in US for four years). When you guys swear 'quaint' seems to come to mind.

So to bring Grammarly back into the picture, I've only ever used it as a secondary, sometimes tertiary spellcheck. It tried to offer its alternative sentences, suggest dialogue changes and when all of that was ignored it offered a trial offer and purchase of the latest gold edition. Finally, it threw a hissy-fucking-fit and only makes the suggestion for the odd full-stop, apostrophe or comma.
I sense begrudgingly though.

So, shipping my stories from MS to Goggle (MS won't accept the particular Grammarly version on their platform) and then shipping it back is wasted time. There are lots of spell checks out there.

 

[Bamagan]

I don't know why that would seem snobbish. I mean by that to write with quirks and idiosyncrasies and changes that would make one's writing seem less programmed. Mix up sentence structure and word choice. Write like one is talking rather than like an essay. I don't know exactly how AI works but I'm guessing it's looking for patterns, so as one writes one can try to avoid patterns that would make one's writing look too smooth or too canned.

I realize that this is a bigger problem than I originally had thought, but it definitely seems to be the case that some people are having a much bigger problem with it than others, and that means that there are things one can do as an author to minimize the problem. It's not snobbish or elitist or dismissive to point that out that this is a problem that to some significant degree probably is within one's control. I have never had this problem with any of my submissions. I also do not ever use AI in my writing.

Tl;dr summary:
Write like a Fremen or get eaten by a sandworm

 

[RejectReality]

I don't know why that would seem snobbish. I mean by that to write with quirks and idiosyncrasies and changes that would make one's writing seem less programmed. Mix up sentence structure and word choice. Write like one is talking rather than like an essay. I don't know exactly how AI works but I'm guessing it's looking for patterns, so as one writes one can try to avoid patterns that would make one's writing look too smooth or too canned.

I realize that this is a bigger problem than I originally had thought, but it definitely seems to be the case that some people are having a much bigger problem with it than others, and that means that there are things one can do as an author to minimize the problem. It's not snobbish or elitist or dismissive to point that out that this is a problem that to some significant degree probably is within one's control. I have never had this problem with any of my submissions. I also do not ever use AI in my writing.

But that's not what you said, which is why I called it out. You said, "Write like a human" which is demeaning as hell.

I read that sentence and all I can see in my mind's eye is Thurston Howell speaking through a perfumed handkerchief.

 

[EmilyMiller]

What is it about AI that makes people all grumpy with each other

 

[PrincessArianrhod]

What is it about AI that makes people all grumpy with each other

It's one thing we definitely, and indefinably, do better.

 

[SimonDoom]

What is it about AI that makes people all grumpy with each other

It's a touchy subject, because people's stories are getting rejected, and while some may be trying to sneak AI stories by I think many are not and are genuinely confused about the process.

I started the other thread to take a more positive spin on it. Instead of trying to guess what the Site is up to, what, based on the information available, can authors do to minimize their chances of rejection?

 

[EmilysRopes]

I don't see 'write like a human' as snobbish but I do see it as about as useful as 'eat like a human'.

Not only do we all have different tastes in food, we all write in subtly different ways. When I worked everyone thought I was extremely good at writing software documentation. Now when I write anything, crueler folk tell me everything I write reads like software documentation!

I do see their point and I suspect when I get round to submitting more stories they may well fail the AI test for being too much like software documentation. Such is life.

 

[StillStunned]

everything I write reads like software documentation!

And of course erotica is closer to hardware documentation...

 

[AG31]

In one of these AI threads in the last couple of days someone posted a list of tips to avoid AI rejection. The were one or two sentences long and included things like "don't accept all of Grammarly's suggestions." I've skimmed a couple of recent threads but can't find it. Can someone point me there?

tia
ag

 

[alohadave]

What is it about AI that makes people all grumpy with each other

I think it's this forum. No matter what you say many someones will take offense to it and write multiple paragraphs to tell you exactly why you are wrong.

There's an eclipse today. Go outside and get some fresh air.

 

[StillStunned]

In one of these AI threads in the last couple of days someone posted a list of tips to avoid AI rejection. The were one or two sentences long and included things like "don't accept all of Grammarly's suggestions." I've skimmed a couple of recent threads but can't find it. Can someone point me there?

tia
ag

Try this one: https://forum.literotica.com/threads/how-to-avoid-ai-rejection.1607906/

 

[AG31]

Try this one: https://forum.literotica.com/threads/how-to-avoid-ai-rejection.1607906/

Thanks much! That's it!

 

[Superfungurl]

THIS IS BASED ON STUFF I HAVE PICKED UP AND IS NOT AN OFFICIAL LITEROTICA ANNOUNCEMENT

I posted this on an AI thread and it was suggested I add it as a stand-alone post. One which might get pinned by @AH_Mod or @Laurel (assuming they agree with the contents).

I have edited it slightly to make it more coherent as a stand-alone message.

March 2024

—

Modified based on further feedback (03/20/2024 17:00 EST)

—

There is a lot of supposition and statements that things are true without any supporting evidence on [AI-related] threads. There is also a lot of emotion. That’s all understandable as I get how much it must hurt to be caught up in this loop.

Having spoken to Laurel (@Portly_Penguin can attest to this) there is absolutely no intent to upset authors. I am absolutely not an official Literotica spokesperson, but I can suggest the following:

1. AI submissions are potentially a big problem for the site. Other sites have become inundated with them, and Literotica wants to avoid this
2. The only person who reviews submissions is Laurel - there is not a team of people doing this, just one overworked woman
3. The AI detection tools are far from perfect, but the site has embraced the principle of zero tolerance for AI. This is to protect authors as much as anything
4. This will inevitably lead to false positives - and the need for additional AI checks has slowed the process
5. Literotica wants to publish your stories, but wants to ensure that they are free of AI. This is for the good of the whole community
6. There has sometimes (and I’m not suggesting anyone on [the various AI-related] threads is guilty of this) been a disconnect between what authors have initially claimed and then later admitted. This has understandably reduced the site’s faith in authors simply stating “I didn’t use AI”
7. This is a very fluid situation and - to employ an over used phrase - an unprecedented one
8. My personal view is that some calm and understanding would go a long way to improving things

Emily

ADDENDUM:

I understand that the problem is mostly using things like Grammarly suggestions to improve text. I’m going to stress this bit. Mostly the author deletes and rewrites the section and it goes through the second time. Mostly… not always (we have heard about exceptions on this thread). So it is helpful to know that at least some stories that were flagged then got approved.

What has been unhelpful (and I’m not suggesting anyone here has done this) is authors accepting that they used a tool with an element of GAI and then submitting the same text saying that they didn’t. So either their first admission was wrong or their second claim is wrong.

Contrary to popular opinion, Laurel neither replies to, nor acts on, everything I contact her about. But she often does and I appreciate it when this happens. Please persist with her, but also give her some time. And if a problem has been identified, please don’t resubmit the same text. This last is maybe contrary to advice provided elsewhere. Either try to address the issue, or get in contact to explain your POV.

For the record as someone who writes professionally I would literally die before going to AI to get something written and pass it as my work...or pass ANY work as my work...other than my work!

 

[Vintage_DM]

This is a hopefully informational addition to this topic, I’m intending to practice what I preach.

I have reviewed I think around eight bits of prose rejected as being AI. The one thing they had in common was a slight uncanny valley formality. Like they had been translated into English. Translated well. But the English was almost too correct. They read more like a user manual. Or - more personally - the type of formal report I write at work.

Maybe that’s a style thing, or maybe the writer has English as a second language.

My writing has errors and colloquialisms and informalities. I use contractions like should’ve and can’t a lot, especially in dialog. I use “gonna” and “kinda” like I do when writing here. I use “ya” instead of “you” and so on. My style is pretty loose, especially with dialog.

Emily

How many times have we seen someone post a comment about a story that has numerous grammatical errors, were the author is roasted for the errors? There is a slang expression here, grammar Nazis.

You are spot on correct, in the course of a dialog colloquialisms and informal comments are frequent. Going on, there are many people here who are not well versed in the fine points of the art of writing. We rely on apps such as Grammarly to act as an editor. Let's be honest with each other. The editor program here doesn't work. I've asked 'editors' for help in the past and never heard a peek back. I'm of the opinion that the site should encompass a wide range of writing styles. Be that a formal business style or a guttural bombast. The language allows for flexibility. People are experimenting with the use of AI. The readers here can and will make a decision on what they like. If the readers see a posting they don't like, the comments flow loudly. What I would suggest is to remove the ability to block comments from being posted about a story. Authors must get feedback about their postings. Comments can and do provide that forum. We've seen where comments become fodder for discussion about a story. That, to me, is healthy.

Y'all keep it between the white lines as you go along da roads.

 

[Melody_Lightdraft2]

THIS IS BASED ON STUFF I HAVE PICKED UP AND IS NOT AN OFFICIAL LITEROTICA ANNOUNCEMENT

I posted this on an AI thread and it was suggested I add it as a stand-alone post. One which might get pinned by @AH_Mod or @Laurel (assuming they agree with the contents).

I have edited it slightly to make it more coherent as a stand-alone message.

March 2024

—

Modified based on further feedback (03/20/2024 17:00 EST)

—

There is a lot of supposition and statements that things are true without any supporting evidence on [AI-related] threads. There is also a lot of emotion. That’s all understandable as I get how much it must hurt to be caught up in this loop.

Having spoken to Laurel (@Portly_Penguin can attest to this) there is absolutely no intent to upset authors. I am absolutely not an official Literotica spokesperson, but I can suggest the following:

1. AI submissions are potentially a big problem for the site. Other sites have become inundated with them, and Literotica wants to avoid this
2. The only person who reviews submissions is Laurel - there is not a team of people doing this, just one overworked woman
3. The AI detection tools are far from perfect, but the site has embraced the principle of zero tolerance for AI. This is to protect authors as much as anything
4. This will inevitably lead to false positives - and the need for additional AI checks has slowed the process
5. Literotica wants to publish your stories, but wants to ensure that they are free of AI. This is for the good of the whole community
6. There has sometimes (and I’m not suggesting anyone on [the various AI-related] threads is guilty of this) been a disconnect between what authors have initially claimed and then later admitted. This has understandably reduced the site’s faith in authors simply stating “I didn’t use AI”
7. This is a very fluid situation and - to employ an over used phrase - an unprecedented one
8. My personal view is that some calm and understanding would go a long way to improving things

Emily

ADDENDUM:

I understand that the problem is mostly using things like Grammarly suggestions to improve text. I’m going to stress this bit. Mostly the author deletes and rewrites the section and it goes through the second time. Mostly… not always (we have heard about exceptions on this thread). So it is helpful to know that at least some stories that were flagged then got approved.

What has been unhelpful (and I’m not suggesting anyone here has done this) is authors accepting that they used a tool with an element of GAI and then submitting the same text saying that they didn’t. So either their first admission was wrong or their second claim is wrong.

Contrary to popular opinion, Laurel neither replies to, nor acts on, everything I contact her about. But she often does and I appreciate it when this happens. Please persist with her, but also give her some time. And if a problem has been identified, please don’t resubmit the same text. This last is maybe contrary to advice provided elsewhere. Either try to address the issue, or get in contact to explain your POV.

Good grief, and I was worried about my stories still in pending after three days. As a self-published author I can understand the danger of generative AI flooding the market with poor quality stories. Harks back to the wild West days of early self-publishing.