People are using AI but not disclosing it. How do you detect it ?

[LurkingInTheDeceit]

This entire conversation just leaves me kinda tired.

Like, yeah, @NumberSix, please don't maliciously sabotage specific tags because you feel spiteful towards a sub-group of people who'd want to use it.

The majority of people who want to use that tag have done nothing to deserve your ire, for one.

Displaying such clear malicious intent is likely one of the few ways I can imagine provoking staff into actually punishing someone for tag use, so you'd be the architect of your own enemy, second.

Sockpuppeting to escape reputational damage or a trail of intent is enough to earn an immediate permaban, third.

Say you accomplished it without our detection/intervention anyway, and you made a bunch of angry people even angrier - maybe some of them are just like you, and decide to retaliate by going after fandoms you care about, or fandoms they dislike if they don't know what you care about, to spread the misery, fourth.

We have an imperfect tag system, but it has utility regardless. I'm working on fixing it up a bit, but compromises between different interest groups are inevitable. At the moment, QQ already has your preferred policy, and doesn't mandate AI tags. Do you really think your spite towards people whose desires aren't even site policy is worth potentially burning the whole tag system down with you?

If so, then there's nothing else to say.

 

[NumberSix]

This entire conversation just leaves me kinda tired.

Like, yeah, @NumberSix, please don't maliciously sabotage specific tags because you feel spiteful towards a sub-group of people who'd want to use it.

Click to shrink...

I was not talking about this place.

If I spam A.I. agent produced content that is mistagged as AI free it will be on AO3 where you can post fics anonymously.

Kindly read the whole conversation for context before you start blaming me for anything.

Frankly the attempt to force mandatory tagging for A.I. works is retarded and should be shut down, and it can easily be bypassed if one is of a mind to do it.
people will not disclose their work, use agentic workflows to make the AI "sound less like AI" and do some basic edits after.
This is a futile, pointless effort.

Might as well add tags for "this was put through autocorrect", "this was written with Grammarly", "this was written on a typewriter" or "this was written via LibreOffice" if you do get around to forcing tags about AI work.

The tools are irrelevant, the end product is the only thing that counts.

 

[LurkingInTheDeceit]

I was not talking about this place.

If I spam A.I. agent produced content that is mistagged as AI free it will be on AO3 where you can post fics anonymously.

Kindly read the whole conversation for context before you start blaming me for anything.

Click to shrink...

Okay, while I'm glad you aren't planning to shit on my porch specifically, most of my points are equally applicable to AO3, as they don't seem to mandate AI tags either.

 

[NumberSix]

Okay, while I'm glad you aren't planning to shit on my porch specifically, most of my points are equally applicable to AO3, as they don't seem to mandate AI tags either.

Click to shrink...

For now they are not.
The anti-AI crowd which is mostly relegated to visual art(read niche porn) would force that shit everywhere though if they had their way.
There is also a pretty severe A.I. derangement syndrome among software devs that start calling all bad code "vibecoded" and some projects have gone so far as banning ALL AI usage irrespective if you are doing minor code modifications or pushing an edited readme through some AI to make the text sound more clear and professional.
There was a similar elitist form of bitching from some of the senile old farts that thought you should not have access to online docs and stackoverflow or use Intellysense back in the day, but right now the bitching is ten times worse.

People are going bonkers and being massive hypocrites about all this.

 

[GatoArc]

For now they are not.
The anti-AI crowd which is mostly relegated to visual art(read niche porn) would force that shit everywhere though if they had their way.
There is also a pretty severe A.I. derangement syndrome among software devs that start calling all bad code "vibecoded" and some projects have gone so far as banning ALL AI usage irrespective if you are doing minor code modifications or pushing an edited readme through some AI to make the text sound more clear and professional.
There was a similar elitist form of bitching from some of the senile old farts that thought you should not have access to online docs and stackoverflow or use Intellysense back in the day, but right now the bitching is ten times worse.

People are going bonkers and being massive hypocrites about all this.

Click to shrink...

And this justifies your behavior? All you managed to do since you started sabotaging the thread was show how deprave pro-AI sentiment could get, honestly I was naive until I met you. The ironic part though? This isn't even an AI hate thread. It's probably the mildest AI thread in the site, yet nope, thats too much for even you. Because you NumberSix know better, and thus as the arbiter of discussion, you deem the thread to be shutdown.

Frankly the attempt to force mandatory tagging for A.I. works is retarded and should be shut down

Click to shrink...

I mean honestly, how is this in any way acceptable behavior?

 

[NumberSix]

I mean honestly, how is this in any way acceptable behavior?

Click to shrink...

You mean like all the people bitching about AI and demanding it be banned/tagged do?
Well, it is better since it does not force people to use tags and force the mods to police that tag usage. As such it is the better approach in this case which is to ignore the demands and do nothing.

 

[GatoArc]

You mean like all the people bitching about AI and demanding it be banned/tagged do?
Well, it is better since it does not force people to use tags and force the mods to police that tag usage. As such it is the better approach in this case which is to ignore the demands and do nothing.

Click to shrink...

No. I mean you purposely going into the thread to sabotage it just because you personally decided that there should be no discussion on the subject. Do not be cutesy by trying to scrub your behavior down the rug.

 

[NumberSix]

No. I mean you purposely going into the thread to sabotage it just because you personally decided that there should be no discussion on the subject. Do not be cutesy by trying to scrub your behavior down the rug.

Click to shrink...

Listing why an idea is bad is now sabotaging it?

 

[doll]

Their entire argument is essentially, "you can't detect AI so what you're doing is pointless" Tell me how that isn't a derailment of the thread.

Click to shrink...

Because it's the entire topic of the thread?
You can't detect AI because AI doesn't exist in the product. It's like asking the ethnicity of the writer. As it turns out, it doesn't matter. No part of the writer's background or process is relevant to the finished product outside of the fact that the final product now exists.

A direct comparison would be if an author knows English as a second language. Sometimes, I'd argue usually, it's very obvious. Very very obvious, and usually to the detriment of the technical merit of the work. Often it doesn't matter because that's not what you're looking for, and when you are usually the level of refinement is such that it's indistinguishable from other products (anyone who is editing AI to such an extent right now to produce actual quality literature really has done pretty every bit of the creative work themselves, although of course matching the quality of most commercial authors is an infinitely lower bar than that). In the end though, the product is the product. The fact that Lolita was written by an ESL author doesn't change the contents of the book, nor does it even really impact that authority of the professor as a literary critic. You're attempting to peer through the words to see how they were made, but fundamentally you can't.

I mean honestly, how is this in any way acceptable behavior?

Click to shrink...

Because mandatory tagging is a settled argument? It doesn't need to be rehashed for seven pages every few weeks because I don't think any one of these threads (which are alike in dignity and character to the 'the wrong type of NSFW is in NSFW and I don't like it' threads i.e. lacking completely) have moved admin opinion on the topic one iota.

Fundamentally, you seem to be fixating on:

That being said, I want to emphasize something, putting aside the reactions towards this behavior, undisclosed AI usage is unethical and it is wrong.

Click to shrink...

Which is both wrong (it's purely a personal, subjective judgement and applies only to your own feelings, not to anybody else (it's also a minority position outside of the broader commercial-artist-fixated bubble which QQ is part of, if that matters)) and perhaps more importantly, is completely off topic. The actual topic is:

How do you prove that a story is AI ?

Click to shrink...

And the answer is: you can't.

It requires information from the author about their process, which they aren't in any way obligated to give you. The only information QQ asks from the author is:
>Is it legal for you to post this (in our jurisdiction)?
Which mostly just concerns plagiarism and locking non-child-safe content behind needing to declare that you are over 18 in making an account. Rule 3 isn't really relevant to AI, so until that changes I don't think you're ever going to see mandatory tagging here. And mandatory tagging is the only way to detect AI, because I could literally get a trained monkey to press random keys on my my keyboard and post that and you would never know. What process a post goes through before it appears on our screens is totally opaque to the readers.

 

[GatoArc]

Listing why an idea is bad is now sabotaging it?

Click to shrink...

OH was that what you did?

The hostile attack on people using AI in this exact thread is proof positive that most people will not admit to using it in fear of ridicule.

Click to shrink...

1st attempt at sabotage.

it will be nice if we can stop with the witch hunting and struggle sessions against A.I. users. The more i think about it the more it smells like a lot of these threads demand people get a free pass for rule 1 violations.

Click to shrink...

2nd attempt

What goes around comes around, the antis have been extremely vociferous and their constant complaints provide zero value,they attack people left and right for liking A.I. content, they are a real pain in the ass on some other sites.

I am just fighting back.

Click to shrink...

The mask comes off.

I'm done entertaining this.

 

[Howldawg]

One way to spot ai is when authors on various sites post either to many stories at once or to many chapters at once though it could be posting a backlog filtering out people who do this has reduced the amount of AI I run into also just spending my time on older fics and checking if the authors are still active as ai is still recent

 

[The Radiance's Follower]

I'm done entertaining this.

Click to shrink...

Lot of bad faith from a lot of people who argue for AI for some reason. They derail, do false comparisons, and project on people here.

Reason why I didn't respond to lightbringer is that he kept bringing "muh job, muh process" And bull crap like that projecting outside discussions upon me while I didn't bring up any of that.

I don't even know why they are here for if all they do is argue for AI. They could talk to their AI chat bots if they needed an echo chamber.

I wish I had the ability to threadban lightgbringer and NumberSix from here. They are disruptive, bring no value, and keep acting like victims.

NumberSix unironically said "I am just fighting back" I am dying man.

 

[Howldawg]

This thread isn't even about weather ai is good or bad just how to avoid wasting our time as readers

 

[TactileFinder]

It depends on "vibes", IE writing styles being repetitive, the narrative of the story seeming simplistic and unoriginal, and what not. People utilizing AI to depict characters and/or their scenes can indicate being alright with AI to an extent, but it isn't a sure fire way of knowing for sure that they're utilizing an LLM for their writing holistically. That said, that's not a terribly useful piece of advice so a more applicable thing to look out for is how often they post.

If they post brand new chapters every day, for several weeks, later in the life of a story... that is often times a signifier of preparing written chapters very quickly. Note, this only applies for stories who aren't just beginning since they could either be crossposting or within a writer's fugue state.

Personally, I just want to read worthwhile stories. AI is often times used as a crutch: to make art that fits your vague specifications then retroactively believe it's what you always wanted as it takes place inside of your imagination, coding that draws upon thousands of scraped examples that you often times either slot in without a care or manually check to make sure it doesn't break anything, or in the case of stories: generating something which fits your vision, story and flow but trying to read over it and make sure it's worthwhile. Is it possible to use AI, and by the end create a worthwhile story? Yes. Is it easier to not do the later and hope the AI can hallucinate it for you? Much more so. This is why I much rather people be accepting of, and normalizing the use of an "AI-assisted" tag because if your story is worthwhile and you openly disclose such things, that means you put thought and care into how your work is read.

 

[ForestDweller]

This thread isn't even about weather ai is good or bad just how to avoid wasting our time as readers

Click to shrink...

I do wish we have a miracle search engine that gives me exactly the fanfics I wanna read.

Face it, most of fics we found are slop we don't wanna read. AI or no AI.

 

[LurkingInTheDeceit]

I wish I had the ability to threadban lightgbringer and NumberSix from here. They are disruptive, bring no value, and keep acting like victims

Click to shrink...

We generally only threadban people if they violate rules, or defy mod directives.

Regardless, you're allowed (and encouraged) to report if a thread is being derailed, and a mod might agree and re-rail it. I'm not gonna get involved in any moderation myself, as I've participated here.

 

[xihute]

It is honestly quite baffling how nasty the ones proclaiming "Witch hunt" every minute of this conversation have gotten.

Click to shrink...

Oh, this is a very good example to show my point!

I don't remember if this was you personally, but some comments were saying, "I don't care if authors are harassed or get nasty comments due to use of AI."

I don't like the attitude, and my point is: what if threads about AI got that kind of engagement you (the ones holding such a stance) are complaining about, and people said "I don't care", "It's your problem", "I'm not doing it", etc.

I wish I had the ability to threadban lightgbringer and NumberSix from here. They are disruptive, bring no value, and keep acting like victims.

Click to shrink...

They did mention why they think proposed solutions for AI detection are harmful or useless.

I would like if we could have normal discussion, without devolving into rants, denunciations, or echo chamber.

If they post brand new chapters every day, for several weeks, later in the life of a story... that is often times a signifier of preparing written chapters very quickly. Note, this only applies for stories who aren't just beginning since they could either be crossposting or within a writer's fugue state.

Click to shrink...

Like with all such methods, it's not 100% foolproof - I've heard about a story that was being posted daily for years, before AI was a thing.

I do wish we have a miracle search engine that gives me exactly the fanfics I wanna read.

Click to shrink...

Yes, this sounds great. "Go go AI, find what I want."