What has happened to this site?

but Grammarly itself claims it does no such thing.
[✔️] Doubt

Without copying directly from Grammarly and inspecting in a hex editor, it would be impossible to prove this.

Two things I can think of off the top of my head:

1. shibboleths in Grammarly's output. Known word groupings that there's a high confidence come from it
2. Use of extended Unicode - the Greek question mark (U+037E ;) is almost indistinguishable from the semicolon (U+003B ; ) - there are many characters like this if you explore alternate languages even in the normal Unicode space.
 
Last edited:
I've never had a problem.

I've also never used anything but basic spellcheck.

You want to use tools to polish your story? Risk it never seeing the light of day here.

Want to self-edit the best you can, you won't have an issue

It's that simple; it really is.
 
After a hiatus of several years, I decided to post a story. Expecting it to be published in 3 to 4 days, I was surprised when 20 days elapsed so I sent in a note. However, rather than prodding publishing, it was rejected with language that insinuated I had used AI. Not a single word in the text was generated using AI. I used Grammarly and MS Word to review spelling, punctuation, and overuse of passive tense as I have done for years. I'm directed to edit and republish the work but I have no idea what to address nor why I need to. I have to say I find it utterly ridiculous and a sign of lazy site management.

Hang in there. Your stories have excellent sexual tension.
 
[✔️] Doubt

Without copying directly from Grammarly and inspecting in a hex editor, it would be impossible to prove this.

Two things I can think of off the top of my head:

1. shibboleths in Grammarly's output. Known word groupings that there's a high confidence come from it
2. Use of extended Unicode - the Greek question mark (U+037E ;) is almost indistinguishable from the semicolon (U+003B ; ) - there are many characters like this if you explore alternate languages even in the normal Unicode space.
#1 is definitely detectable, though of course it has a degree of false positives.

#2 is another variant of Unicode watermarking, and like you said, we'd need to look at the byte output of Grammarly to see it. But it's not hard to do at all, by anyone really, so if Grammarly was indeed lying about their practices, then I suspect we'd have heard about it by now. (Especially since there are benefits to leaving AI-generated watermarked, since it can help you exclude it from later scraping & training and thus minimize the feedback-loop pollution of your model).
 
Just saying what I see (as attached for any naysayers)

I reiterate I have never used Grammarly, yet I still got this recent rejection note.

My work is my own yet after over 100+ uploads I received a similar rejection for the perceived use of AI in a tale that published after resubmission and no edits

So in summary some use Grammarly, others do not, but irrespective Lits pre publishing sift detects it or false detects it and no one can give a definitive answer in absence of official word from moderators in chief

I refer then to the original question raised by the OP

With a backlog of completed stories parked (that for continuity's sake cannot post until my purgatory story sees the light of day) I’m down to just hoping the issues will be fixed… forgetting all about the sooner or the later.
 

Attachments

  • IMG_0267.png
    IMG_0267.png
    267.4 KB · Views: 13
Last edited:
What's the technical mechanism that Grammarly might use to leave behind traces that it has been used? Assuming you submit your story as text in the submission form, not as a Word file or something, since a proprietary format can hide all sorts of metadata we're not aware of.

The only way I can think of is Unicode watermarking using unprintable characters, but Grammarly itself claims it does no such thing.
Yeah, I have my doubts about it as well. I always submit a Word file so I suppose it can preserve some artifacts, but I admit I'm clueless as to why would copy pasting into submission form textbox keep those artifacts.

I guess we speculate a lot here, hoping that sooner or later we might actually guess something right. Word groupings and patterns seem like a more logical explanation to me, too.
 
I reiterate I have never used Grammarly, yet I still got this recent rejection note.
Did you use one of the other tools mentioned?

My work is my own yet after over 100+ uploads I received a similar rejection for the perceived us of AI in a tale that published after resubmission and no edits
A false positive.

So in summary some use Grammarly, others do not, but irrespective Lits pre publishing sift detects
We cannot be sure that Grammarly/other is always detected. Either it is not always spotted, or maybe it is but low use is tolerated.

With a backlog of completed stories parked (that for continuity's sake cannot post until my purgatory story sees the light of day) I’m down to just hoping the issues will be fixed… forgetting all about the sooner or the later.
I feel for you particularly on this, as I have been in a similar position.
 
This will really really annoy you, but in the last fifteen minutes I've had a PM exchange with Laurel about a story edit glitch - we've exchanged PMs about a fix. All in less than an Oz day, what's more.

I suspect your problem is, you're in the wrong time zone. You need to be tomorrow, like me.

Carry on ;).
I am glad to here that she responds to you as well.

Maybe I won't be viewed as the unicorn in the forum any longer.
 
This will really really annoy you, but in the last fifteen minutes I've had a PM exchange with Laurel about a story edit glitch - we've exchanged PMs about a fix. All in less than an Oz day, what's more.

I suspect your problem is, you're in the wrong time zone. You need to be tomorrow, like me.

Carry on ;).
Nah, I suspect Laurel keeps track of her followers and fanboys, and you simply rank high.

Simon, on the other hand... I remember him not getting a reply the last time he PM-ed. Is he just a low tier in Lit's pyramid? :ROFLMAO:
 
We cannot be sure that Grammarly/other is always detected. Either it is not always spotted, or maybe it is but low use is tolerated.
Almost every AI detection tool I have seen reports back on the "percentage" of the reviewed text that is suspect of being AI generated.

50% suspect AI generated from a 10K word story is only 10% of a 50K word story. This is something that people may want to keep in mind when deciding how much they want to break their story up before submitting. We don't know what the threshold might be with the detection process Lit uses, but keeping below it would likely minimize rejections.
 
Lits pre publishing sift detects it or false detects it and no one can give a definitive answer in absence of official word from moderators in chief
I have never heard of anyone here receiving a rejection specifically for the use of Grammarly. There are a lot of assumptions that its incorrect use can get a story flagged for AI use, but as you have said yourself, you don't use it and a story can still get flagged.

I, on the other hand, use it every time and have never had an issue with it. It's like using a hammer; hold it right and use it correctly and the head hits the nail. Screw around with it and your thumb feels the pain.
 
I've never had a problem.

I've also never used anything but basic spellcheck.

You want to use tools to polish your story? Risk it never seeing the light of day here.

Want to self-edit the best you can, you won't have an issue

It's that simple; it really is.
I keep reading posts that basically say that AI is how things are done now and that Literotica should just shut up and accept it. My answer to those posts is something I learned a long time ago.

Just because you can doesn't mean you should.

Like with any other technology, AI should be a tool to be used when applicable, not a tool to be used just because it's there and it's convenient, or because as I suspect in some instances, because it's the latest cool thing to do.

I use Word and the "Review" ribbon is no doubt powered to some extent by AI. I use it only to check for spelling errors and punctuation. Letting it reword my sentences is no different from having a human editor edit my stories in a way that words the story as the editor thinks it should be written instead of what I wrote.
 
Just because you can doesn't mean you should.
Exactly, but I think we are currently at the stage of learning what AI is good for, and what it isn't.

A lot of technologies have had particular uses tried that did not catch on (steam-powered cars, anyone?).
 
I think we are getting sidetracked by talking only about grammarly here.

Sentence and phrase autocomplete at the micro level (one sentence at a time) can take a person who sincerely didn’t use ai to envision-create-compose their overall story, and nonetheless fill it with n number of individual sentences that were ai completed.

And it could be: your computer or device overall, your word processor, your web browser itself, the particular website you may be visiting, and sneaky “turns it back on with each update” policies. It’s an ongoing cat and mouse game to keep all of it turned off.

Furthermore, more and more people use the web versions of word. Turning it off there is I believe another thing altogether. So we also need to adjust to asking in exacting and painful detail what version and how they used whatever editor they used, including “web version or installed/client/workstation version”.

If OP says they used free grammarly, I believe them. You have to be patient and have a high tolerance for ads to put up with its incessant attempts to upsell its paid version.

Let’s switch gears and talk auto-completion for phrases and sentences.
 
Last edited:
Back
Top