I tested a highly rated AI detection tool, it is worthless

[chloehunt]

I had two stories kicked back for AI use on my alt account. I was shocked and furious because I don't use AI to write my stories. I use Grammarly to check for typos and that's all. I don't let it rewrite anything. I paid for a subscription to Originality.ai's content scanner and did some tests. I scanned my story once, and it came back as 13% likely AI written. Completely false because I know for damn sure I didn't use AI to write any part of it. I scanned it again after fixing some line spacing that messed up during the copy/paste, and it reported 45% AI written???? I scanned it a 3rd time, and now it's at 25%??? F*cking scam software much?

Some writers have a high false positive because we use the rules of English and write formally, or the software is just hyped up sh*t. First person writing seems to cause false positives more, and that's one of my favorite ways to write about a sexual experience! Now I'm having a New Year meltdown because writing is a huge part of my life. I've been writing original stories for over ten years, and now software is claiming IT ISN"T MY MINE??? Is this what I have to look forward to for the rest of my life? Competing to prove I'm not a f*cking robot? I lost interest in painting because AI took that over, and now my writing is being steamrolled. I plan to cancel my Grammarly subscription. I wonder if they are feeding our data to these crap detector tools. Anyone not using Grammarly getting false positives? My content before I started using Grammarly has less false positives. Can anything be done about this before i lose my f8cking mind? Anyone else make progress overcoming this?

 

[ElectricBlue]

Can anything be done about this before i lose my f8cking mind? Anyone else make progress overcoming this?

There's a bunch of threads over in the Authors' Hangout which gather up the latest information. Which isn't much, other than reports from other frustrated folk like yourself.

 

[PandorasBachs]

I had two stories kicked back for AI use on my alt account. I was shocked and furious because I don't use AI to write my stories. I use Grammarly to check for typos and that's all. I don't let it rewrite anything. I paid for a subscription to Originality.ai's content scanner and did some tests. I scanned my story once, and it came back as 13% likely AI written. Completely false because I know for damn sure I didn't use AI to write any part of it. I scanned it again after fixing some line spacing that messed up during the copy/paste, and it reported 45% AI written???? I scanned it a 3rd time, and now it's at 25%??? F*cking scam software much?

Some writers have a high false positive because we use the rules of English and write formally, or the software is just hyped up sh*t. First person writing seems to cause false positives more, and that's one of my favorite ways to write about a sexual experience! Now I'm having a New Year meltdown because writing is a huge part of my life. I've been writing original stories for over ten years, and now software is claiming IT ISN"T MY MINE??? Is this what I have to look forward to for the rest of my life? Competing to prove I'm not a f*cking robot? I lost interest in painting because AI took that over, and now my writing is being steamrolled. I plan to cancel my Grammarly subscription. I wonder if they are feeding our data to these crap detector tools. Anyone not using Grammarly getting false positives? My content before I started using Grammarly has less false positives. Can anything be done about this before i lose my f8cking mind? Anyone else make progress overcoming this?

AI detection software is beyond useless at this point.

 

[darkride]

I asked Chat GPT to write an episode of a tv show. I was curious what it would look like - how good could it really be? I gave it some particular parameters, and on its first attempt (which is as far as I went), it produced a story worth around 7 minutes of screen time (instead of 23 mins), which had no real substance / plot. So... basically what I was writing when I was 8 years old.

Pretty scary, to be frank...

As far as detecting AI-created content? My understanding is there is no way to accurately perform this task. I've heard of several universities that were using AI software to scan assignments that were being submitted by students, and they were getting a VERY high false-positive result. One lecturer decided to scan one of his own papers, which he knew he had written, and the AI said it was fake too.

I haven't had any pieces knocked back anywhere for being AI, but the way I write - generally very A, B, C - I expect it's only a matter of time...

 

[aToyEnjoyer]

Anyone not using Grammarly getting false positives? My content before I started using Grammarly has less false positives. Can anything be done about this before i lose my f8cking mind? Anyone else make progress overcoming this?

I am not using Grammarly or any other AI or even non-AI grammar assist tool whatsoever. The only assist I have is Google Docs' spellcheck, I even turned off the grammar assist in there too since it was worthless most of the time and just annoying me.

Yet part 3 of my story still got falsely flagged for AI writing anyway. The only advice in the rejection messages and FAQs is to stop using AI assist programs, so wtf am I supposed to do?

 

[cadeauxxx]

I asked Chat GPT to write an episode of a tv show. I was curious what it would look like - how good could it really be? I gave it some particular parameters, and on its first attempt (which is as far as I went), it produced a story worth around 7 minutes of screen time (instead of 23 mins), which had no real substance / plot. So... basically what I was writing when I was 8 years old.

I tried to argue this in another thread yesterday, only to get told that Chat GPT is apparently, very good at writing stuff. At this point, there isn't much of an attempt to listen to us who have suffered, and we're just going in circles arguing about what AI is. There are many people who clearly don't know what AI really is, and will quickly point the finger at those of us suffering from false flags. The goal posts are constantly shifted in these arguments and nothing is resolved.

There needs to be a general agreement in these arguments about AI being something that generates text based on prompts. Spellcheck and grammar tools are not meant for generating from prompts. And the ones that can (Grammarly has a setup for it), pose the question of how do you even prove someone is using this?

I am not using Grammarly or any other AI or even non-AI grammar assist tool whatsoever. The only assist I have is Google Docs' spellcheck, I even turned off the grammar assist in there too since it was worthless most of the time and just annoying me.

I have pointed this out too in another thread; only to get told that apparently, word suggestions in these programs are now AI. Even though Microsoft Office Assistant came out in the early 2000s and browsers have built-in grammar and spell check software that has existed for many years before the AI boom.

I'm afraid we're stuck in limbo here until or unless the site owners figure out how useless the AI detection systems are. Since AI is trained on human writing, it's only natural that it's going to come out trying to mimic a human in the first place. This is clearly a moral panic, and we're the unfortunate victims caught in the crossfire of it.

 

[aToyEnjoyer]

I have pointed this out too in another thread; only to get told that apparently, word suggestions in these programs are now AI. Even though Microsoft Office Assistant came out in the early 2000s and browsers have built-in grammar and spell check software that has existed for many years before the AI boom.

Man, the extent of what the spell check does for me is correct a word from "gaurd" to "guard" if I spell it wrong by mistake. I'm not even using word suggestions.

I'm afraid we're stuck in limbo here until or unless the site owners figure out how useless the AI detection systems are. Since AI is trained on human writing, it's only natural that it's going to come out trying to mimic a human in the first place.

The least they could do is GLANCE at a story before rejecting it because the system said "85% AI" or whatever. Fully relying upon these horrible systems to reject stories is insane, I'm certain my story would be easily accepted upon any sort of manual review.

 

[cadeauxxx]

Man, the extent of what the spell check does for me is correct a word from "gaurd" to "guard" if I spell it wrong by mistake. I'm not even using word suggestions.

I have the same problem lol. Spellcheck always messes up names for me.

The least they could do is GLANCE at a story before rejecting it because the system said "85% AI" or whatever. Fully relying upon these horrible systems to reject stories is insane, I'm certain my story would be easily accepted upon any sort of manual review.

I feel the same way.

I'm the one who posted on here days ago my message that my story was "ran through multiple systems and came back a high percentage of AI generated text".

These detection tools are worthless since AI is trained entirely based on human writing. It's meant to mimic someone else's work, which would be a real person and not a robot.

I believe there are probably AI generators out there that are being trained entirely based on stories on this site. Literotica is a huge site with hundreds of thousands of stories. In the past, there was the big problem with authors having their work plagiarized and published on Amazon Kindle. I was one of the many who had their work stolen and published on there. It would not shock me at all that someone has taken bulks of stories on here and ran it through AI, and now the true authors are being flagged after a detection tool has a database built on our writing styles.

 

[aToyEnjoyer]

I'm the one who posted on here days ago my message that my story was "ran through multiple systems and came back a high percentage of AI generated text".

That sucks so goddamn much, even makes it sound like they trust the AI detection programs more than their own eyes. AI IS NOT TRUSTWORTHY. Begging and praying this doesn't happen to me, I'm only on my first resubmission.

I have the same problem lol. Spellcheck always messes up names for me.

Add to dictionary, add to dictionary, add to dictionary...

I believe there are probably AI generators out there that are being trained entirely based on stories on this site. Literotica is a huge site with hundreds of thousands of stories. It would not shock me at all that someone has taken bulks of stories on here and ran it through AI, and now the true authors are being flagged after a detection tool has a database built on our writing styles.

Almost undoubtedly, although I'm quite new to Lit so that wouldn't even apply to my own personal style.

In the past, there was the big problem with authors having their work plagiarized and published on Amazon Kindle. I was one of the many who had their work stolen and published on there.

At least the content of my stories so far wouldn't allow for them to be published on Amazon, lol

 

[Britva415]

As far as detecting AI-created content? My understanding is there is no way to accurately perform this task

Not with AI, anyway.

Humans can suss it out because the AI-generated material ultimately doesn't make sense, and contains errors no human would make.

But what human has the time for all that.

 

[aToyEnjoyer]

The proper course of action here is to combine the AI detection with a manual review. It's what some game companies are doing with anti-cheat and what discord is doing with their new disallowed image detection.

The AI system detects possible hits, then a human follows up to review. The AI is nowhere near trustworthy enough to run on its own, but it does speed the process up immeasurably.

However, in this case, I seriously doubt the capability of the AI writing detection programs. To the point where the above process would just be so full of false positives it wouldn't even be worthwhile anyway. Which makes it even more bizarre to me that Literotica is putting so much blind faith into them, but also makes me kind of understand why they aren't doing manual reviews?

 

[LonRivers]

I think the problem is that some people use programs and internet websites to check their grammar and spelling. Whenever I read a post about a rejected story, Grammarly is mentioned. I think the reliability of these sites and programs are for shit. If my grammar is not the best in some areas or if I misspelled a few words, oh well. I try to keep an eye on that when I am writing. I am not going to upload my stories to anything buy Lit.

BUT I also think that if a story is rejected with AI concerns, the person who rejected it SHOULD explain just how or why they feel an AI generator was used. I understand that this can be time consuming, but if you are going to accuse an author of something, you should be ale to explain why.

 

[LonRivers]

I am not using Grammarly or any other AI or even non-AI grammar assist tool whatsoever. The only assist I have is Google Docs' spellcheck, I even turned off the grammar assist in there too since it was worthless most of the time and just annoying me.

Yet part 3 of my story still got falsely flagged for AI writing anyway. The only advice in the rejection messages and FAQs is to stop using AI assist programs, so wtf am I supposed to do?

I use LibreOffice and the built in spellcheck only. if I think a sentence doesn't look right, I will rephrase it or rewrite the line all together so that I am happy with it. Grammar be damned sometimes.

 

[forgotmyantidepressants]

Let me start off by saying that I am very passionate about AI and what its use is currently doing to the artistic community as whole, not just Lit authors, and my language reflects that.

Sapling AI detector is also a joke. I ran a story that was rejected for AI generated content through it and it came back "0% fake". Fake. What an ugly four letter word to describe a story/creative project that a person probably put their heart and soul into. I have a few choice four letter words for people who feed the works of others into ChatGTP to churn out  fake, soulless garbage in an attempt to make a quick buck (some authors are suing ChatGTP. As they should).

I've been experimenting with first person present tense. A 1.3k word short was graded, "100% fake".

These detectors are obviously crap. The frustrating thing about running stories through them is that we are in turn training the AI. We are making it better. We're honestly shooting ourselves in the foot. I think Literotica is doing this because they don't want to be sued. I get that, and I support their efforts in promoting stories written by humans. But I think they are being overzealous in rejecting stories with little to no explanation.

 

[forgotmyantidepressants]

I had two stories kicked back for AI use on my alt account. I was shocked and furious because I don't use AI to write my stories. I use Grammarly to check for typos and that's all. I don't let it rewrite anything. I paid for a subscription to Originality.ai's content scanner and did some tests. I scanned my story once, and it came back as 13% likely AI written. Completely false because I know for damn sure I didn't use AI to write any part of it. I scanned it again after fixing some line spacing that messed up during the copy/paste, and it reported 45% AI written???? I scanned it a 3rd time, and now it's at 25%??? F*cking scam software much?

Some writers have a high false positive because we use the rules of English and write formally, or the software is just hyped up sh*t. First person writing seems to cause false positives more, and that's one of my favorite ways to write about a sexual experience! Now I'm having a New Year meltdown because writing is a huge part of my life. I've been writing original stories for over ten years, and now software is claiming IT ISN"T MY MINE??? Is this what I have to look forward to for the rest of my life? Competing to prove I'm not a f*cking robot? I lost interest in painting because AI took that over, and now my writing is being steamrolled. I plan to cancel my Grammarly subscription. I wonder if they are feeding our data to these crap detector tools. Anyone not using Grammarly getting false positives? My content before I started using Grammarly has less false positives. Can anything be done about this before i lose my f8cking mind? Anyone else make progress overcoming this?

Please keep painting! Don't let these tech bros with no conscience or a shred of creativity keep you from doing what you love!

Consumers need to come together and refuse to engage with "art" made by AI. People who use AI don't care about art. They care about money. If we refuse to give to them, they'll leave us in peace.

 

[chloehunt]

I think the problem is that some people use programs and internet websites to check their grammar and spelling. Whenever I read a post about a rejected story, Grammarly is mentioned. I think the reliability of these sites and programs are for shit. If my grammar is not the best in some areas or if I misspelled a few words, oh well. I try to keep an eye on that when I am writing. I am not going to upload my stories to anything buy Lit.

BUT I also think that if a story is rejected with AI concerns, the person who rejected it SHOULD explain just how or why they feel an AI generator was used. I understand that this can be time consuming, but if you are going to accuse an author of something, you should be ale to explain why.

Yes, I fully agree.

The purchased software I'm testing changes its conclucion if you change the paragraph spacing before you scan the text. The text and story are still the same, but the AI detection rate jumped from 13% to 47% when the paragraph spacing was corrected. You may as well be flipping a coin if the software doesn't actually consider the content and formatting of the story. It appears to be looking for patterns with no context. A human has to read it and make the call, or your just flipping a coin and throwing out false accusations.

 

[chloehunt]

Let me start off by saying that I am very passionate about AI and what its use is currently doing to the artistic community as whole, not just Lit authors, and my language reflects that.

Sapling AI detector is also a joke. I ran a story that was rejected for AI generated content through it and it came back "0% fake". Fake. What an ugly four letter word to describe a story/creative project that a person probably put their heart and soul into. I have a few choice four letter words for people who feed the works of others into ChatGTP to churn out  fake, soulless garbage in an attempt to make a quick buck (some authors are suing ChatGTP. As they should).

I've been experimenting with first person present tense. A 1.3k word short was graded, "100% fake".

These detectors are obviously crap. The frustrating thing about running stories through them is that we are in turn training the AI. We are making it better. We're honestly shooting ourselves in the foot. I think Literotica is doing this because they don't want to be sued. I get that, and I support their efforts in promoting stories written by humans. But I think they are being overzealous in rejecting stories with little to no explanation.

Yes, i've shed tears over this. I feel like I have to film myself being creative now, and even that can be faked.

 

[aToyEnjoyer]

I've been experimenting with first person present tense. A 1.3k word short was graded, "100% fake".

Wondering if my writing in third person present tense could also contribute to being falsely detected as AI. I know it's a fairly strange way to write, it could be that most real writers are worse at it than the AI.

The purchased software I'm testing changes its conclucion if you change the paragraph spacing before you scan the text. The text and story are still the same, but the AI detection rate jumped from 13% to 47% when the paragraph spacing was corrected.

Right, because paragraph spacing is simple enough that AI can get it right more often than real humans. But this just punishes the human writers that are the most competent and had their work stolen by the AI in the first place!

 

[Nutbutterbish]

Did you resubmit? They encouraged me to change it before I resubmit. But. Hell, I wrote the shit like I wanted.

 

[astuffedshirt_perv]

These detectors are obviously crap. The frustrating thing about running stories through them is that...[Lit is] being overzealous in rejecting stories with little to no explanation.

fixed the last paragraph for you

 

[forgotmyantidepressants]

fixed the last paragraph for you

Thank you.

 

[chloehunt]

Did you resubmit? They encouraged me to change it before I resubmit. But. Hell, I wrote the shit like I wanted.

I did. It sat there for a week, so I took it down. My motivation has fizzled, but I'll attempt to rewrite it and run it through the AI detectors again. There has to be a style that doesn't cause false positives so often. (Yay, extra work...) Different line spacing in the same story causes fasle positives to go up by 30%, so it may be pointless.

 

[aToyEnjoyer]

It sat there for a week, so I took it down.

My resubmission has been pending since Jan 4th now, I'm leaving it up until something happens and I'll reply here with the result.

 

[Nutbutterbish]

I did. It sat there for a week, so I took it down. My motivation has fizzled, but I'll attempt to rewrite it and run it through the AI detectors again. There has to be a style that doesn't cause false positives so often. (Yay, extra work...) Different line spacing in the same story causes fasle positives to go up by 30%, so it may be pointless.

Damn, homie. You are so passionate, I really hope you can get back to the grind. That peculiar rejection is really frustrating, though, extremely when awaiting that approval for so long. My anxiety is always through the roof when I post. I hate waiting.

 

[Portly_Penguin]

-