AI Allegations Thread

[astuffedshirt_perv]

I don't know whether anyone's been following the news, but there seems to have been a breakthrough in legislating AI in the EU.

And also on 11 Dec, Mixtral was released, open source, roughly ChatGPT 3.5 effectiveness. The cat has done left the barn.

 

[PandorasBachs]

So...my story that spent two weeks in pending before being sent back as AI (it's a story I wrote 7 years ago and had published here previously under a different username) is now going to take another asinine amount of time to go through the resubmittal process?

Whatever they're using as AI detection software is complete and utter BS as none of the existing software is even remotely accurate. Laurel and Manu need to be open and honest about what's going on, because they're the ones screwing things up right now.

 

[dmallord]

Out of curiosity, I used one AI checker MillyDynamite listed. My 500-word limit was an intro into a story I wrote. It came back: 0% Chance that AI generates your text. I guess I can't even imitate an AI writer.

I had one rejection for AI attempting to use pictures/AI generated for scenes in a story. I ended up scraping the eye candy elements and published the story as text - the usual stuff. I had set a personal goal of entering a story in each category, but my stick figures [the best ones I could draw] just looked like some scribbled on a bathroom wall. I gave it up - the idea of posting pictures in a story.

 

[MillieDynamite]

Don't blame me, @dmallord! Just kidding, you can blame for anything you want to blame me for.

Out of curiosity, I used one AI checker MillyDynamite listed. My 500-word limit was an intro into a story I wrote. It came back: 0% Chance that AI generates your text. I guess I can't even imitate an AI writer.

I had one rejection for AI attempting to use pictures/AI generated for scenes in a story. I ended up scraping the eye candy elements and published the story as text - the usual stuff. I had set a personal goal of entering a story in each category, but my stick figures [the best ones I could draw] just looked like some scribbled on a bathroom wall. I gave it up - the idea of posting pictures in a story.

 

[ShelbyDawn57]

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.

Going to call you out on this. having someone or something call out grammatical errors like comma splices and run on sentences is essential to improving your skills. My mind immediately went to my high school English comp teacher and what she would do to my rough drafts. Marking them with that dastardly red pen of hers, pointing out all the things I did wrong. How else are we supposed to learn if not through the guidance of others. Do you use editors or beta readers? Same concept.

PWA, Grammarly, they are decent tools and, used judiciously, can help make you a better writer by making you aware of things you do repeatedly. Of course, if you blindly take their advice, or any other sources advice, with out reserve, you are abdicating your responsibility as the author of the work.

 

[DDX84]

My mind immediately went to my high school English comp teacher and what she would do to my rough drafts. Marking them with that dastardly red pen of hers, pointing out all the things I did wrong.

In that example, your teacher didn't make the changes for you. It's not the editor's job to come up with a huge list of changes and then integrate them into your document on your behalf.

Of course, if you blindly take their advice, or any other sources advice, with out reserve, you are abdicating your responsibility as the author of the work.

I agree, excellently put. An abdication of responsibility. You've nailed it.

 

[ShelbyDawn57]

In that example, your teacher didn't make the changes for you. It's not the editor's job to come up with a huge list of changes and then integrate them into your document on your behalf.



I agree, excellently put. An abdication of responsibility. You've nailed it.

Yup, It's not using the tool, but how you use it that is the issue. Just like so many other tools.

 

[MillieDynamite]

Actually, using Word, your editor makes changes and sends the altered file back to you. You either accept the change or reject it depending on your intent in the individual sentence. In the old days, they used blue pens to highlight mistakes, and since the document would be double-spaced, they made notes of what the correction should be. Again, it is up to you to change it or leave it alone. Sometimes they simply write clunky sentences and or paragraphs in the margins way back when.

In that example, your teacher didn't make the changes for you. It's not the editor's job to come up with a huge list of changes and then integrate them into your document on your behalf.



I agree, excellently put. An abdication of responsibility. You've nailed it.

 

[AG31]

I tried to find a/the thread is that most generic about AI issues. Did I succeed?

Today I went and read older stories by people who complained as being falsely rejected. I was sobered to find that some of these stories seemed to me to be klunky enough to have been generated by a non-human. I don't really read a lot of stories here on lit, but it's possible that bunches and bunches are just not well enough written to be clearly human generated.

And AI is going to get better.

So maybe we should just throw in the towel and dedicate a forum to AI generated stories. If they're good and fun to read, it would be a feature, not a bug. It wouldn't exactly solve the problem of people falsely claiming AI stories as their own, but it would reduce it.

I think someone has already suggested this idea, but this thread has 27 pages, and I couldn't think of a word to search for. If you did suggest this, would you speak up???

 

[djrip]

I tried to find a/the thread is that most generic about AI issues. Did I succeed?

Today I went and read older stories by people who complained as being falsely rejected. I was sobered to find that some of these stories seemed to me to be klunky enough to have been generated by a non-human. I don't really read a lot of stories here on lit, but it's possible that bunches and bunches are just not well enough written to be clearly human generated.

And AI is going to get better.

So maybe we should just throw in the towel and dedicate a forum to AI generated stories. If they're good and fun to read, it would be a feature, not a bug. It wouldn't exactly solve the problem of people falsely claiming AI stories as their own, but it would reduce it.

I think someone has already suggested this idea, but this thread has 27 pages, and I couldn't think of a word to search for. If you did suggest this, would you speak up???

Poor writing, in a technical sense, is not an indication of AI; just the opposite. AI writing is technically impeccable.

 

[AG31]

Poor writing, in a technical sense, is not an indication of AI; just the opposite. AI writing is technically impeccable.

Yeah, I didn't see any technical problems. Just klunkiness.

 

[djrip]

Yeah, I didn't see any technical problems. Just klunkiness.

I'd like to know which ones you mean and what is meant by clunkiness. AI writing I've seen has a particular style - technically perfect, written like the best student in the class followed the structure guidelines to the letter. Uncreative, yes. Clunky? I would not describe it that way.

 

[ShelbyDawn57]

I would be interested in knowing the page length for stories getting tagged. My assumption would be those writing short, one or two page chanters are more likely to get flagged than those writing longer complete stories, but it's just speculation.

 

[AG31]

I'd like to know which ones you mean

WritingWhatILike was one of the authors I looked at. Clearly a lot of people like them. And I don't know what "AI authored" writing is supposed to be like. But it struck me as klunky. No technical problems.

 

[djrip]

WritingWhatILike was one of the authors I looked at. Clearly a lot of people like them. And I don't know what "AI authored" writing is supposed to be like. But it struck me as klunky. No technical problems.

I'll check them out too.

 

[djrip]

WritingWhatILike was one of the authors I looked at. Clearly a lot of people like them. And I don't know what "AI authored" writing is supposed to be like. But it struck me as klunky. No technical problems.

I don't know. I just read the first half page and it doesn't seem anything like AI to me at all. It doesn't sound like the teacher's pet writing a five paragraph essay exactly as specified, repeating the main idea of each paragraph in the final sentence and summarizing at the end. It has paragraphs that jump around from idea to idea haphazardly, the way a human thinks.

 

[AG31]

I don't know. I just read the first half page and it doesn't seem anything like AI to me at all. It doesn't sound like the teacher's pet writing a five paragraph essay exactly as specified, repeating the main idea of each paragraph in the final sentence and summarizing at the end. It has paragraphs that jump around from idea to idea haphazardly, the way a human thinks.

Which story did you look at? I looked at the first one in the list.
But this does raise the issue that was bugging me when I posted this thread. I'm not recommending you get involved in it. It doesn't seem to be going anywhere. But I would like to see examples of stuff that is getting flagged.

 

[djrip]

Which story did you look at? I looked at the first one in the list.
But this does raise the issue that was bugging me when I posted this thread. I'm not recommending you get involved in it. It doesn't seem to be going anywhere. But I would like to see examples of stuff that is getting flagged.

Sorry, I looked at ch 1 of the story he said ch 2 got rejected. Her Son's Surrogate.

 

[djrip]

Which story did you look at? I looked at the first one in the list.
But this does raise the issue that was bugging me when I posted this thread. I'm not recommending you get involved in it. It doesn't seem to be going anywhere. But I would like to see examples of stuff that is getting flagged.

Yeah the first one, A good Sleep, doesn't seem at all AIish to me either. Dialog mixed in, variety of tags and sentence structures, etc, not super repetitive, etc.

 

[AG31]

Yeah the first one, A good Sleep, doesn't seem at all AIish to me either. Dialog mixed in, variety of tags and sentence structures, etc, not super repetitive, etc.

Can you point me to an example of an AI generated bit of erotica? ChatGPT won't go beyond a couple of paragraphs for me.

 

[djrip]

Can you point me to an example of an AI generated bit of erotica? ChatGPT won't go beyond a couple of paragraphs for me.

I'm just comparing to what I've been able to get chat gpt to produce when messing around with it, which has been extremely formulaic. I'm sure it can be made to produce other stuff, but if I'm going to reject something for AI, I want to be able to read it and think, yeah that seems like it has some characteristics of AI.

 

[AG31]

I'm just comparing to what I've been able to get chat gpt to produce when messing around with it, which has been extremely formulaic. I'm sure it can be made to produce other stuff, but if I'm going to reject something for AI, I want to be able to read it and think, yeah that seems like it has some characteristics of AI.

How do you identify "formulaic?" Is it different than "klunky?" In the story I read there was one sentence after another with the same structure.

Note: I got inspired to go on this quest when someone else in the writingwhatIlike thread said he could see how it might be flagged.

 

[ShelbyDawn57]

How do you identify "formulaic?" Is it different than "klunky?" In the story I read there was one sentence after another with the same structure.

Note: I got inspired to go on this quest when someone else in the writingwhatIlike thread said he could see how it might be flagged.

When I think of formulaic, I am reminded of the street I grew up in. There were two or three very nice houses, say 4k sq. ft., on big lots and then a couple dozen 2K sq. ft. tract homes. They were all pier and beam, wood frame construction, but just from looking, you could tell the difference. The larger custom homes had style. Thought and effort went into their design, the finishing touches were nicer. The tract houses all pretty much looked the same. Many were exactly the same floor plan.

Perhaps that's what is being tagged as AI, simplistic sentences structure, limited vocabulary, repetitive style, overused word choices. Not casting any stones here as I haven't read much of anything the authors that have complained have written, just offering a thought.

I advance my style by reading, and not just here. So many times I've encountered something and just stopped and said, "Wow, I love how they did this or that." and then tried to emulate it in my writing. Not plagiarizing, just enhancing style by seeing what works for others and, conversely, noting what I don't like.

I find myself leaning towards the answer being, it's all about style when it cones to AI. But I'm just onw writer...

 

[djrip]

How do you identify "formulaic?" Is it different than "klunky?" In the story I read there was one sentence after another with the same structure.

Note: I got inspired to go on this quest when someone else in the writingwhatIlike thread said he could see how it might be flagged.

I just asked chat gpt to write a story with dialogue. First para identifies lily and Oliver, known for their adventurous spirits and boundless imaginations. And uses once upon a time.

First dialogue para - Lily, with fiery red hair and a freckled face, pointed toward a path. "Hey Oliver, bla bla."

Next one - Oliver, with his unruly mop of brown hair and twinkle in his blue eyes, grinned. "Nope, bla bla."


Just, very simple and formulaic, meaning, seeming like it follows a script. Character name, with their hair characteristic and one other description, verbed. "Hey other character name, bla bla."

A good sleep sounds nothing like that.

 

[mypenname3000]

I would be interested in knowing the page length for stories getting tagged. My assumption would be those writing short, one or two page chanters are more likely to get flagged than those writing longer complete stories, but it's just speculation.

Both mine were in the 7k size using free version of Grammarly to find basic mistakes like missed question marks, a mess up in verb tense, agreement issue. That's it.