"AI" Rejection

[djrip]

I'm also not hearing anyone claim they were unfairly rejected because they never used any AI. Feels like a couple of the thread participants have actually outed themselves by their own admissions. Their complaint seems to be 'I want to use more AI than you allow'.

People getting rejected as AI when they did not is exactly what the thread is about and what happened to most of the people complaining. Maybe one or two were arguing that AI assisted rewriting shouldn't be prohibited.

 

[jhealy55]

You know, if the rejection would have said "Dude your dialogues are so bad, they could be written by AI," I'd have a thing to follow up on. But basically, that would be a quality gate, not an AI gate. It would be a lot easier to accept.

Overall, the stance seems to be that AI writes bad and boring texts and that LIT doesn't want that. Why not just reject every story below 3.5 after 2 weeks? That would actually improve that site a lot.

Your writing in this thread doesn't seem at all clunky or difficult to follow, so I don't think you are that far off in getting that excerpt to flow more naturally than it currently is.

Whether it's AI or quality of dialogue that's causing the rejection, I don't think you missed an opportunity to connect with a ton of future followers with this submission. Not sure that's any consolation, but I think most readers would have given up pretty quickly. It's very tough to connect to either character and immerse yourself in the scene as it is currently written.

But I've written way worse, I can assure you. It's a process.

 

[jhealy55]

People getting rejected as AI when they did not is exactly what the thread is about and what happened to most of the people complaining. Maybe one or two were arguing that AI assisted rewriting shouldn't be prohibited.

and again. I came in looking to support writers in that circumstance. that's not what I'm seeing argued through most of this thread

 

[Littletoo]

My last 3 have been rejected on "ai" grounds . One of which has been on my harddrive for at least 5 years (a started story that never got finished) long before Ai writing was a "thing" I added a note to say I have the original proof file if it's needed, resubmitted, and it got rejected 2 more times...if my next post (which I know for a given, hasn't been anywhere near an ai generator) gets rejected, I'm seriously thinking of not submitting anything else

 

[Littletoo]

holy crap you wrote this in the middle of November and then it happened here last week!!! You're like Nostradamus or something. You seriously need to play the Powerball

I don't know why the Literoticon Mods keep looking for AI.

Instead of looking for Artificial Intelligence, they should be looking for the Real Thing.

The other thing they should watch for is MYT( Monkeys With Typewriters). My roommate is a math professor and she can prove that MWT, if you're not in a rush, will reproduce the entire body of written works, even literature that hasn't been written yet . But she says this doesn't work if the monkeys have intelligence, artificial or not. They must type absolutely randomly or the math doesn't converge. They can't even have memory, because then the processes are Markovian, and again, the proof doesn't converge.

My roommate tried to explain Convergence in Probability to me. She got as far as Lebesque Measure before I fell asleep. Then I had nightmares about being sexually assaulted by monkeys on dirtbikes Too much TV news I guess. Sometimes I think I'm on the Wrong Planet.

 

[NoTalentHack]

Your writing in this thread doesn't seem at all clunky or difficult to follow, so I don't think you are that far off in getting that excerpt to flow more naturally than it currently is.

Whether it's AI or quality of dialogue that's causing the rejection, I don't think you missed an opportunity to connect with a ton of future followers with this submission. Not sure that's any consolation, but I think most readers would have given up pretty quickly. It's very tough to connect to either character and immerse yourself in the scene as it is currently written.

But I've written way worse, I can assure you. It's a process.

Okay, how about this?

This is from the middle of my second story on the site, currently sitting at 4.67 in Loving Wives. The only tool I used to write it was Google Docs, which has a spell-checker; there's a grammar checker, too, but it's no more effective than Word's, i.e., it's mostly a pile of crap.



According to Sapling, it's 100% AI generated. In reality, 0% of it is. If the site were using Sapling, I'd be fucked. More than that, back when I wrote it, I'd have no idea where to go to fix it. Knowing what I know now, could I do better? Sure. It's very basic, and I recognize that. But, again, amateur writing site, and a very early work from me.

Now, we don't know what detector Laurel is using, so let's run it through a few more.

GPTZero claims 24%:



ZeroGPT:



CopyLeaks says it's human:


Detecting-AI:


Content@Scale says it "passes for human," which really gives the game away, doesn't it?:


Scribblr:


GPTKit:


Eight different tools. Eight different assessments, varying wildly between "this is completely human" to "this is 100% AI." Most of them that find "fake" text can't even agree with each other which parts are fake and which are real! And given that we don't know which tool Laurel's using, we can't even give advice on how to make someone's text less likely to trip it.

AI detection companies are a scam. They're all selling something, whether that's "peace of mind" or tools to evade content detection. If the only thing they do is to sell AI detection (and sometimes even if they sell other things as well), they have caps on how much data a free user can put through them, so they're not useful for determining if a text is AI generated as a whole. As an example, here's another length of text from the same story passed through Sapling:


If I, as a new writer, had pasted the whole story in, it would have grabbed that one section and said, "A-OK!" because it would have cut off the rest that it was offended by. For the record, the witch-hunter had the paid version, and he said (although I think he might have just been a troll) that my story was flagged as 81%.

AI detection is a scam. Period. End of story. Anyone who tells you otherwise is ignorant or selling something. And if you don't want to hear that from me? How about a few machine learning experts?

https://arstechnica.com/information...-think-the-us-constitution-was-written-by-ai/

(which has one of my favorite AI detector images)



Don't Use AI Detectors For Anything Important

Apparently I Am A Robot

 

[alohadave]

It's official? Can we pee freely in the pool now?

Every pees in the pool, but if you pee into the pool, they will insist that you leave.

 

[secondlullaby]

Alright. I've caught up with the entirety of this thread because I was interested in it. I'm a little disappointed at how it's devolved near its end, because I think a number of really interesting subjects arose earlier in the conversation.

This. We need to get clarity on the trigger phrases and/or sentence structures, so we can compile the differences between human writing and derivative word prediction software (let's not call it AI, because it isn't).

I think we can all recognise AI when we read it, but reducing it to the "reasons why" does get a little harder.

I wanted to mention here: A lot of word prediction software does exactly the same thing as full-on chatGPT style AI, just at a much smaller scale. It functions based on pools of data to predict sentences. If the pool of data is based on your own text, it's still AI.
Also, I think we can't recognize AI easily in smaller samples, as evidenced by the next post I'll make below (went over character limit)

Two of the principal things AI detectors are doing when examining a text are checking for what is being called 'perplexity' and 'burstiness.'

Thank you for this info. Burstiness is always something I've striven for but didn't have a name for.

I guess that is just how I write. I actually use a couple of websites with lists of different works for said, as well as sexual words, so that things are varied. I try to find the ones that feel right in those moments. I like to make them seem cute. Maybe I shouldn't. Most of my readers seem to like what and how I write.

Personally I look for a different approach to the sentence structure (rather than just replacing a word) when my text shows signs of too much repetition. It's more difficult and it can require reworking of whole paragraphs, but for me it's worthwhile.
I know a few people who write with a thesaurus close at hand, and results vary widely. In some cases it's far, far too obvious.

I will never, ever use Grammarly, Scrivener, et. al. Conspiracy theory maybe, but their algorithms...

You mentioned you use an older version of Word, and that makes me think you're exactly the kind of person who might love Scrivener. It's not an AI software at all. It's a fairly simple text editing program with much better organizational tools than anything else I've used. Give their trial a try, and you'll see that it isn't what you think.

Electronic calculators were predicted to inevitably lead to no one being able to do math, and spell-checkers would mean no one could spell. Going back further, radio was the demise of books and newspapers, recordable cassette tapes the end of the music industry, and so on.

It's all part of the hype cycle that happens every time there's some technical innovation. Excitement is matched by dire predictions and pearl-clutching overcorrections... the perceived need for AI detectors and companies rushing them to market despite their inaccuracy part of that.

This is also my take on the subject, but I'd thought of AI (in the context of visual art as a clearer example) as more comparable to something like handmade pottery. There's a pretty clear advantage to mass production of dishes, which traditionally took time and artistry to make. But the upside is, potters still do make money, because people still want the real/artistic version of the thing everyone buys mass market versions of now.

So I do actually think this threatens creatives in a sense, because we will have to change our ways of doing things and remarket ourselves. We may lose parts of the industries we've targeted. But I do think there will be a market for our work, and supply and demand dictates that those serious about it will still do well, albeit on a more competitive basis.

I don't want to have to change the style of my writing because some AI system can't tell the difference. I've never used AI, have the editing on Google Docs to prove it, and just use some grammar checkers for minor issues - although that in itself seems to be the issue which is absolute insanity. For it to be suggested that I use AI is an insult despite the numerous insistences to the contrary.

If an AI system can't tell the difference, it does say something about your work. AI functions, as far as I know, based on medians. If your work looks like AI, then it probably doesn't stand out as anything other than average.

I'm surprised it doesn't make you want to take a closer look at your writing and make changes, because that's part of becoming a better writer.

Do you need to be a better writer? I guess not. That's up to you.

(Please note, I have no opinion about your writing skills and haven't read your stories. I just mean, I would be bothered if my work was flagged as AI, because to me that would suggest I had written something average and boring, which isn't what I strive for)

Well said. I do have sympathy with Laurel... quite a bit less with those who say "just stop writing badly lol".

I don't think everyone needs to strive to be better at writing (or anything else)... but I do think it's fine to weed stuff off a website that way, if you're in charge. Not that I think that's what Laurel is doing. Just that we're all subject to her decisions, and I'm fine with that.

Writing should have the goal of hitting a reading level between 6th grade and 8th grade. Just because you graduate high school or college, doesn't mean you want what you read for pleasure to be post graduate level. In fact, many people with high educations, don't want to concentrate on a story like it is a technical manual on brain surgery.

Why 6th-8th? That's the average American reading level, which means it alienates a lot (up to half) of your readers.

I ran my story "Take Me Home with You" through a few readability score websites, and depending on the index got a Grade 2-5 reading level, with a Flesch Reading Ease score of 90. This is exactly what I strive for in my work, because it's accessible to nearly everybody, but also because ease of reading lets everyone enjoy the story (well, not everyone will enjoy that particular story, but you know what I mean).

This is kind of what you're saying already, but, I think it's worth aiming for the lowest reading level you can (without sacrificing story-- and there's where the skill lies). A lower reading level does not have to mean you're writing "I am Sam. Sam I am," and it does not imply a certain type of content. It speaks to general readability, which is appreciated even by those with high-level reading skills.

https://arstechnica.com/information...-think-the-us-constitution-was-written-by-ai/

Most of them use a combination of checking for "perplexity" and "burstiness."

The bottom line, from the article, is this:

I wrote a bunch on this (and why they mostly suck), my own process, the witch hunt issue, and more over at https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-26#post-97985369 and in later parts of that thread, mostly in the process of disassembling a witchhunter's arguments.

Some other key posts:

My process, which utilizes tools in a way that Laurel has no problems with, and which has never gotten dinged as AI: https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-19#post-97969898

Non-writing places where AI detection/assessment has failed horribly: https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-24#post-97976494

A primer on how ChatGPT works in the first place: https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-24#post-97979855

A quick assessment of the tool the witchhunter was trying to say "proved" that I and others used AI text generators, flagging all but one of the top 10 most read stories on the site (the newest of which was posted in 2009): https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-28#post-97991991

My thoughts on the hypocrisy of "purists" saying we had to conform to their preferred tools: https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-28#post-97992782

And, over in another thread, the way that I did use AI in artwork, how I did, and the things I'm still wrestling with: https://forum.literotica.com/threads/about-that-ai-assist-in-writing.1600310/post-97980469

I'm probably going to turn all of this into an essay at some point. Loving AI, here I come!

I only put this quote in here so I wouldn't forget about it while going through the thread. Please do write the essay. I love this stuff.

 

[secondlullaby]

I’d seen her a few times at our local nightclub; I only visited the intimate small town establishment on Thursday nights, when the DJ played rock music, and only when my work rota permitted it. The young brown-haired lady I watched from the raised bar area was present on every occasion I frequented the basement nightspot. But she was out of my league; she exuded confidence and sexuality, dressed to attract attention, and had the erotic power of a porn star. I was a 33-year-old slightly overweight divorcee with the start of a receding hairline.

The first couple of times I saw her, I was just one of another 300 revellers in the nightclub. The time after that, we spoke. I bought her a drink at the bar, and we chatted above the thumping beat of AC/DC and the nostalgic melodies of Green Day.

Then just last week, she teased me; we got a table at the rear of the club, next to the coat check, and beside the bar. The alcohol helped; I always found it difficult to speak to new people, but we conversed easily, and her sex appeal and flirtatious behaviour encouraged me to invest a small chunk of my bonus money in her inebriation. She coquettishly intimated and exaggerated, hiking her short skirt higher when our drinks ran dry, before I staggered to buy replenishments.

Khristyna blew kisses with her soft red lips, and she daintily rubbed the back of my hand as she discussed sex. I’d evaded the subject and had not mentioned my proclivities. My divorce was still raw, and I had found it difficult to find women who had similar perverted kinks in the small Cheshire town. Instead, I imagined the scenes of depravity that the coquettish Lithuanian described as she revealed her fantasies and experiences. However, as my alcohol consumption rose, I shared more of my sexual desires.

She giggled as she saw my expression, adjusted her tight top and suggested we get another round of drinks. Clearly, I was on a promise.

Actually, she had tormented me. The Eastern European immigrant had seduced and then discarded me, kissing me goodbye as she staggered from the nightclub to “go home” to her flat. She promised to text me as she shimmied out of the club, wiggling her bum in her incredibly tight skirt before she ascended the stairs.

For days, she dominated my sex-heavy dreams; I fantasised about running my lips over her naked bosom and parting her legs to explore her wondrous womanhood. Probably waxed or shaved, hairless to not impede my view. Each night, I visualised the ravishment of the cheeky sex bomb, and I woke to a painful erection.

I discussed my predicament with my friend the day before the next nightclub session; the assistant nurse was an accomplished womaniser and had the phone numbers of the shadiest contacts in the small town. “Twenty five quid,” he suggested. “It’s a drug that loosens inhibitions. She’ll be gagging for it. Just slip it into her drink and bone her!”

It sounded unethical and illegal. A medical consultant should not entertain the clandestine application of drugs to a sexual partner, and yet the following day I dressed to visit the nightclub with a small sealed paper packet in my pocket, intending to tip the contents into Khristyna’s drink. I knew she’d be there; she messaged me.

That Thursday night, I arrived as the doors opened, bagging the table in the far corner. The dark venue, lit by strobe lighting across the nightclub, left a few of the tables near the coat check in gloomy shadow.

Khristyna bounced over to me, dressed in a shiny black short skirt with a tight bustier and dark stockings. She grinned when she saw me and flounced onto the chair opposite. “Hiya Joe. I’ve had a crap day.”

“Drink?” I asked, and walked to the bar to buy her favoured tipple - a vodka and lemonade. My hands shook as I picked the hi-ball glass; there was no way for me to empty the contents of the paper packet in my pocket without a dozen witnesses seeing it.

As the night wore on, and Khristyna flirted more explicitly and drunkenly with me, my mind whirred as I struggled to find a way of getting the contents of the inch square paper sachet into her drink.

Under the cover of the club darkness, I ripped the top of the sachet in my right hand and gestured towards the stage. “Hey, isn’t that guy waving at you?” With one smooth motion, I reached for my beer on the table, emptying the white crystals into her glass.

My heart pounded. There were over 300 potential witnesses to what I had done, but as I waited for the shouts and challenges, no-one uttered a word. “What guy?”

“Oh, he’s gone now.”

The dark-haired beauty chuckled, and I hurried to the toilet to calm my nerves before returning to the table to buy us both another drink. But the polluted tipple worked; she kissed me, sliding her dainty hands over my navy shirt.

She tasted divine; the harsh bitterness of the lemons on her tongue with the sensual beauty of the Lithuanian kitten. “Let’s go,” she whispered, nibbling on my ear. I couldn’t wait to depart, stumbling into the cool evening with my new sexual conquest. She opened the car door of a parked red vehicle nearby. “Taxi?”

The hooded driver grunted. “Yeah, where to?”

“17 Roseberry Gardens,” I replied, scooting in the back seat with the sexual powerhouse. Her hands rubbed my thighs as she pushed her lips onto mine, and the ten-year-old car lurched forwards. She could not stop touching me; I tugged at her tight rubberised corset, exposing her large breasts on her lissom frame to my fingers.

Touching her nipples sent a frisson of excitement to my cock as I rolled the points between my fingers. She pushed her lips onto mine as the car sped along the streets, braking gently at our destination. “We’re here,” the driver muttered, reaching into my pocket to get my wallet and pay our cabbie.

Hands grabbed my wrists, pulling them away from my body towards the front seat. Two clicks signified she had cuffed me in one smooth motion. Khristyna smirked as she readjusted her corset and picked my wallet from the floor.

“You can’t do this,” I spat. They had set me up to be mugged, and I looked out of the window; this was not my house, but a desolate industrial estate.

Notes moved to next post, ran out of characters.

 

[secondlullaby]

(continued from my post above, replying to @bawdybloke )

So. This is text I get bored of quickly, and I can see why it would be flagged as AI. I know I'm being blunt and harsh, and I'm sorry to do it publicly. I've kept the entire quote here and have spent a lot of time on this to give you some very specific reasons why, in the hope that it might help you or someone else.

The first thing that stood out to me about your writing is the overuse of adjectives and alternate nouns. We don't need a different way of saying every term in the story. It makes our focus drift around, and makes your descriptions less powerful (even the much better ones in between).

AI seems likely to do this in a less discriminating way than it should, because of how it functions. Here, you've not been as sparing as you could be. You refer to Khristyna in so many different ways that it is confusing, and I have to re-read sentences just to make sure we're still talking about the same person.

In many places, the prose is pedantic or just too wordy. I don't think that's always a bad thing--You are, after all, writing in a first person perspective, and perhaps your main character is a pedant. But if that's the case, the focus should be on how he comes across. I think it's too hard for readers to identify with someone using words like "inebriated" inside their own thought processes. Let him be a pedant in his speech, if he is, but skip it in the descriptions. Otherwise you're the pedant, and you will lose readers that way.

I think this is a common "style" for a lot of introverts who enjoy reading, especially us analytical types. We know the words, and we want to get out all of the details in an efficient way. But in adding all those details, we actually take away from the larger story and make it more difficult to read. AI does the same thing.

I've realised while going through this that the blue text is actually just a specific example of this problem.

As the text progresses through this excerpt, I noticed a shift that I found interesting. It becomes less wordy, but more confusing. You're skipping over details in your descriptions that are important. I get that the car sped along and then stopped, but I don't get the feeling of it happening. Why did the driver reach into our main character's pockets? Why the focus on hands grabbing his wrists, while keeping them anonymous? Why not show the thought process of his realisation he was set up? Why use the word "mugged" when there must be more to this than the wallet?

To me it seems like the descriptions earlier in the story are your remedy to these problems. Maybe they do solve the problem, but they create a new problem.

If you're interested in trying out some changes to your writing, maybe a writing exercise like this one would be a good starting place. I'm not saying efficiency is everything, but learning how to be efficient when you want to could add some variation and interest.

EDIT: I lost all the highlights on the original post and had to re-do them, and probably missed things or highlighted a little differently. On re-read I've noticed, you really try to fit in adjectives at every possible opportunity. The problem with this is, it's just too much info. There's hardly a noun in this prose without some kind of description attached to it. If you focus on the most important ones, you'll get a lot more impact with your words. I think with some editing-- to cull the unnecessary stuff-- I might quite enjoy this story.

 

[secondlullaby]

I actually did go through and change some small suggestions, but that was primarily due to how they sounded to me, not that they sounded AI. They were overly formal for the type of story because I'm used to writing Historical so my word choices can be slightly more extravagant and didn't fit the style for what I was posting.

Considering my most recent part - the one that was rejected for AI but posted after a month - is sitting at a 4.97 rating, I'd say it's above average. Maybe don't make assumptions based purely on a message in a forum without having read the work in question.

I'm not sure if the strikethrough was a retraction. In case it wasn't: I was not making assumptions about your stories. I was specifically avoiding it.
EDIT: I've skimmed over bits of "Light The Fire Pt. 03" and don't see any issues with your writing at all. It doesn't "pop" necessarily, but it doesn't need to. You write effectively.

As for the usefulness of ratings: I think ratings aren't based entirely (or even mostly) on writing quality. I actually go looking for highly-rated but poorly-written stories here, because I have a deep respect for the ability of some writers to tell a story.

I'm not a naturally good storyteller (and I attribute my ratings being less than I'd like to that). The people who have better ratings than me but lesser technical skills are exactly the people I need to learn from.

In turn, maybe I could share some writing skills. Or maybe that's my ego talking.

 

[ElectricBlue]

So yeah as you showed, a lot of stuff is gonna get flagged even if it's been posted here for decades

Only if some clown goes around reporting stories, who can't read in what year the story was published.

The defence against an AI accusation, if someone thought of doing something stupid like that, is, "Duh, Chat GPT hadn't been released when that story was published, numb nuts."

I can't see the site screening old stories - why would it do that? That's a daft thing to suggest.

 

[ElectricBlue]

I'm not suggesting the site would do it, I am saying that if it did all kinds of stuff would be flagged even if it's impossible the author used ai. The fact that these crappy detectors are calling stuff from 2006 ai generated shows they ain't good for shit so no one should be using them

I responded to what you actually wrote:

a lot of stuff is gonna get flagged even if it's been posted here for decades

More precision in your sentences, then .

 

[NoTalentHack]

I responded to what you actually wrote:

More precision in your sentences, then .

Some dude literally came through one of the other threads claiming that the level of output was a reason to ping someone for maybe being AI. Another poster mentioned me as a counterargument. This dude, after a couple more exchanges, then proceeded to hateread, run through Sapling, and report every one of my stories over a two day period. Nothing's come of it so far--because fuck that dude and Sapling--but do I think someone would do this out of spite? Yeah, I do. I'm not saying Laurel would pull people's stuff, but I imagine she'd at least run it through a detector. And, as discussed, my stuff, especially the early thing, at least trips some of them. I imagine a lot of other new writers' work would, too.

 

[RejectReality]

I can't see the site screening old stories - why would it do that? That's a daft thing to suggest.

I would hope they did. The only real way they could possibly determine that whatever tool they're using has any degree of accuracy would be to test it on the existing catalog of stories that couldn't possibly have used AI.

As seen in the examples here, many of them fail spectacularly.

If it wasn't tested against exactly what gets submitted here regularly, that could explain why there are so many false positives.

 

[ElectricBlue]

I would hope they did. The only real way they could possibly determine that whatever tool they're using has any degree of accuracy would be to test it on the existing catalog of stories that couldn't possibly have used AI.

I don't think that would prove much. There's so much content here, as we've already seen, both old and new, that "looks like" AI generated content, which means the only solution would be to throw any any detectors, and rely on a human's judgement. Which is precisely what started this debate in the first place. People didn't like Laurel's judgement.

I don't know what the solution is - but nothing I'm reading in any of these threads is a "solution". Using AI tools to catch AI, without a truckload more sophistication, doesn't cut it. I think we all agree on that.

 

[NoTalentHack]

Using AI tools to catch AI, without a truckload more sophistication, doesn't cut it. I think we all agree on that.

The problem is that some people don't. They're confident their stuff won't get tagged by it, so they're willing to roll the dice.

 

[RejectReality]

I think the highest AI-generated rating I've come across on any of mine was somewhere in the 30% range. 1% is a false positive, but at least the way I write doesn't seem to trigger these things much.

It would be almost hilarious if the first chapter of the one I have in editing now gets pinged because I wrote it in 2019. LOL ( At the latest. That's just the earliest date I could track down when I went looking ) Finished 9 last year but only just recently summoned up sufficient energy and motivation to start editing. Sapling says 0% on the first 2000 or whatever number of words it checks for free.

I suppose the indignation we're seeing could be similar to what we see with underage rejections and such where a little digging determines that the prohibited content did in fact exist, but the poster didn't consider it to be such. People may not even realize that whatever tools they're using are spitting out AI generated suggestions.

Doesn't really feel like that, though.

 

[secondlullaby]

Telling people they should change their writing style based on unreliable detectors is not only extremely rude but also disturbingly patronizing.

I’m not suggesting anyone “should” change their writing style. This person seemed to be asking why their post was flagged as AI and I gave my opinion on why that might be.

It was my hope to make a constructive comment, and I spent quite a while trying to pinpoint what changes might be helpful for the writer. I’d appreciate if someone did the same for me, especially if they went to the effort to give specific examples in my work, and pointed to outside resources.

 

[secondlullaby]

Pacifying a machine by changing text is worse than letting it generate its own. The rules forbid AI rewrites, but pinpointing its preferred edits essentially allows it to rewrite anyway.

I don’t care about pacifying the machine, though. I care about improving my writing so it’s better than the machine, which I would think would make it more clear I’m not a machine, as a side effect.

I could definitely be wrong.

 

[secondlullaby]

you said you ran your stories through readability score websites. How about you run them through the Sapling ai detector? I'm real interested to see what it says

This looks like it might be sarcasm?

In any case, I was curious, and results for my noncon story ranged from 0-22% “fake” depending which part of the story I put in. This actually makes a lot of sense to me because the beginning of the story is written more generically than I’d like and gives the highest result.

I only put in four sections of the story to test, and 2000 characters isn’t a lot. There may be higher results elsewhere in the story.



An alternate AI detector which someone else mentioned gave me a “100% human written” score, which is accurate but of course it does show there’s variation in how they detect AI.

I don’t think anyone should strive to write “like me,” as an aside (you said something like that earlier). I think we should all strive to make our own writing better.

I have a lot of work and practice to do in the storytelling side of writing, but I’ve been writing professionally for years and have had my work published several times. I’d like to think I can be helpful in some areas I have confidence in; e.g. sentence structure.

That’s some ego talking, yes. But my intention was to answer a question that was asked, in a helpful manner. I regret that it came across differently.

 

[secondlullaby]

And how would he do that? He also can't know how the text was created in the first place. So, you just let someone else make the gut decision. "Oh, you should try to get Wilma as editor, she isn't so hardcore on AI sounding text..."

That was my gut reaction too, but I think @Dybbuk was talking about AI editing software, before. An editor would know the edits were less likely to be AI generated if they had suggested to the writer themselves.

Anyway, to me it’s all kind of irrelevant. We’re on Laurel’s platform and she makes the rules. I’m fine with that.

I’m just enjoying keeping up to date (ish) about how AI writing works and how it differs at the moment from human writing.

 

[TheArsonist]

My final point is that if you want to make a change to how the site submission rules work, you are going to need to get a lot of open-minded and persuade-able authors to come over to your side and take up your cause. I was fully both before I started to engage on this thread. Everything I've seen so far has pushed me farther into the other camp.

I understand that a confrontational response can harden opposition, but I think you should also recognize that your own contributions to the thread have been pretty inflammatory. You've acted very dismissive about authors getting trapped in a Kafkaesque process of false and insulting accusations that they have no way of defending against, and being arbitrarily blocked from the site on those unfair grounds. Despite all the evidence that AI detectors are junk that regularly (not just in rare cases, but frequently) classify human-written texts as AI-generated, you completely ignore that and persist in arguing as if they can generally be trusted. You've suggested that people who disagree with you must be arguing in bad faith, surreptitiously using or wanting to use AI more than they admit. And like many of those defending the policy, you go pretty far towards implying that even if authors haven't actually used AI, it's no great loss and they deserve to be kicked out anyway because their writing must not be good enough.

Do you not see how that leads to heated responses?

You've written some of my favorite stories on this site, so I'm sorry to find you on the other side of this controversy. I hope you reconsider.

 

[theWollstonecraftWoman]

The temperature seems to be coming down a little in here.

 

[MillieDynamite]

Really @theWollstonecraftWoman, you think it's calmer in here?

The temperature seems to be coming down a little in here.