Hang On

[dlodoski]

Hey Guys,

First of all, I want to assure everyone that this forum is for fun and expression.

But, at the same time, we're going to insist that contributors display a certain amount of respect and courtesy to copyright holders. In an ideal world, AI prompters would ask permission first. But insisting on that is not realistic.

However, the rules that will be in place for this forum *are* realistic, and will be enforced. Comments are welcome.

[Duncan Edwards]

For anyone who might not get it, let's examine this work of "art". The original watermark is missing and has been replaced with one from DALL-E. That's a big give away right there. Krystal is original real. The legs on the right are not.

6ydpce7b6qre1.jpeg

While I don't get the attraction to anime or a lot of the pseudo-anime cartoon stuff, we all have our own groove. Giving credit for the original work is the proper way to do things and will certainly help keep you in the good graces of the people who can delete your stuff. If you don't know where it came from you can at least offer that.

[Viridian]

I am in accordance with the stance against using copyrighted material. It's disrespectful to the original creators to have their work - which they put the effort to create - put through an AI filter to generate the same content in a form they did not approve of.

I lost my shit when someone took my AI work and turned it into anime and published it. Work that I originally made using an anime generator and turned into photorealism, then turned back into anime by someone else under the notion of "look how cool this is!" If I wanted an anime version of my own work, I'd make it myself.

I'd be in favour of a more hardline stance on banning the use of copyrighted source images, especially from other creators in the QS community. Get permission, period. It takes zero effort to flip a photo into an anime style. It's an insult to both the original creators who put the cost and labour into creator the original work, and other AI creators who spend time composing original images.

I don't want to make the rules sound like it encourages people to experiment with others' work. It's almost as easy to make your own original work with AI. We AI creators generally don't touch each other's work, and this rule set mostly came about because someone without AI art experienced faffed around with a ChatGPT image generator.

If we can't enforce it, I'd like to see something that encourages transformative use of original work instead of lazy style switches that are literally done with a single button press.

[dlodoski]

These points are well taken. And completely justified/understandable.

My thoughts when creating the rules (short list for now, but likely to grow) were guided by pragmatism. People are going to play with copyrighted content regardless of what stance I take here at QSF. And that's because there are certain 'elements' that exist in the producer catalogs (and some mainstream as well) that AI in general hasn't learned yet.

I was waiting for something to cross a line before taking some action. Anyone here remember the politician who famously said "I can't define obscenity, but I know it when I see it"? It was something like that.

Long and short - play nice and everybody is happy.


Footnote - one of the reasons we're being a bit permissive here is the idea that some of the producer spawned AI output is worthy. But having said that, inane amounts of output gets boring after a while. Unlike naval antiaircraft batteries in WWII, quantity does not take on a quality of its own. It cheapens it.

[Viridian]

I would still like to push for a harder line on the use of copyrighted content without permission. At minimum, especially concerning the use material made by our own community, there should be an ask-first policy.

True, people are going to play around with copyrighted material. There's no one stopping that. But no one is required to publish that creation. We already disallow reposting of original works. There's nothing less pragmatic about requiring consent to publish AI work derived from copyrighted source material.

I feel the rule stating that you can use others' creations as long as credit is given is openly endorsing theft. AI is already heaviliy stigmatised, so insinuating that you don't even have to ask to use someone else's work as a base is a step backwards.

If creators are open to the idea of seeing their content transformed with AI tools, that is an individual choice, that can either be declared on the creator forums or stated in specific cases where an AI creator seeks permission. In other words, creators should voluntarily opt-in, not the other way around.

This is closer in line to what artists already do. We don't take each other's original characters and works and republish them. I only make Safari Girl renders because Acidtester has endorsed them. I create tributes to DerMudGecko and QSA because they are long gone and only with significant transformative work inspired by their work, not using their work. I have permission to publish graphic-novel renders of Micheal Leach's photography - but I only intended to show him what could be done with his work, not as a public creation. I wouldn't feel comfortable using Duncan's photography, especially knowing that he is already using his vast library for his own AI work.

In other media, people ask if they use my characters. aikofell doesn't just slap Quicky Sanders on a playing card and give credit. I don't run aiko's 3D models through AI without proposing the idea first. That is the standard level of courtesy and the rule should not assume that this isn't viable to enforce.

Not to mention that lots of creators are vehemently anti-AI and do not want their work sullied by artificial intelligence.

Let's not make this a guessing game as to who is OK with their work being used and transformed in what manner and to what degree. It's so easy for anyone to use AI at a basic level that the current rules are an open invitation to turn this forum into an artificial Wild West.

I really want to underline that what triggered this is barely AI art. It's literally uploading an image to ChatGPT and typing a single sentence. There is no actual individual creative effort. You can literally do this yourself, right now, in about two minutes (source obviously QSV).

Screenshot 2025-04-04 at 17-42-00 Ludella Sinks! #3 Ludella sinks into your waking fantasies even if you're on a budget! - UMD.png

ChatGPT Image Apr 4, 2025, 05_51_26 PM.png

[Duncan Edwards]

I wouldn't feel comfortable using Duncan's photography, especially knowing that he is already using his vast library for his own AI work.
...

I really want to underline that what triggered this is barely AI art.

I don't really have any good answers to the questions here. Not completely certain what all questions are for that matter. I do know that one thing I have discovered, and I've mentioned this to Dave, is using existing material to create with is often more difficult and produces a less satisfactory result than just going for an AI original. It's actually quite limiting because AI is already way ahead of anything I've done. Most just don't seem to know how to ask for it.

Duncan_Edwards_cinematic_shot,_Anita_Ekberg_dressed_as_a_Mongol_p_2370de98-b145-4689-8af3-ba6e57c389b5_upscaled.png

Duncan_Edwards_cinematic_shot,_Anita_Ekberg_dressed_as_a_Mongol_p_72968b18-713c-45de-93bb-251a4f1f938b_upscaled.png

Duncan_Edwards_cinematic_shot,_Anita_Ekberg,_mouth_open_with_the_2d233fe3-bd14-47e8-9e1a-203bef430ea0_upscaled.png

There's some playing I did with how Anita Ekberg should have appeared in The Mongols. I didn't need anything from MPV. I just took it from the aggregated collection of everything on the web, or that is, the AI did. We could discuss that meaning for a few days and still not have any real firm answers. I just know that now our work has become part of that everything. Do I owe Anita Ekberg? The film producer? The creators of the folk legends it's based on? Where do we draw the line?

Another thought rolling around in my head goes something like this. The total work that goes into creating a simple original photo is so much greater by comparison to AI that the effort in even sophisticated AI remains trivial. I look at any of it and can't see any dirt, sweat, expense, travel, all that very real, tactile, stuff. My thinking says any AI you claim as your own is itself just another blender full of "the everything".

I know where Ludella came from as well as every element in the photo and the equipment to take the photo and the locations and Ms. Lodoski who cooked breakfast for us that morning, all of it. If somebody makes it look like something it's not, what do we do? I don't care much for all the drawings and animated appearance or anime in general. Comparing the real thing with a digital sketch just doesn't bother me that much no matter how simple it is to do. If anything I care even less.

We view all of these things from our own points of reference and that's mine. I don't have a lot of firm answers yet.

[dlodoski]

Ok, so another rule has been added.

This was has to do with flooding topics with large numbers of similar images.

While one or two images may have the desired impact, ten or twenty won't. This fits neatly into the Law of Diminishing Marginal Returns. You don't have to be an MBA to understand the concept of eating ten ice cream cones on a hot day - where the first one or two are great, and you want to vomit by the time you finish.

In any case, this rule is reactionary. And any rules added in the future are likely to be as well. It's the nature of something that's new.

[Viridian]

I also agree with this rule.

Guys, AI is fucking easy to use. The result may look fantastic in that you can create something beyond your imagination with a few words. It's getting even easier now - you don't even have to prompt. Just type it out in ChatGPT, or run an image through an AI filter, and you've got art.

You can make an infinite number of images on exactly what you want You don't have to share every single one. AI images aren't a display of skill - they're not stories that you wrote yourself, or artwork that you learned how to draw. Once you have your workflow - and for some it is just typing in a prompt - you've "mastered" that skill. Spamming 20 images a day on the forum doesn't show anything new. We can all do it, or learn how to do it in 15 minutes.

Many of us go through a honeymoon period when we first get into AI. The beauty of AI is that you can, indeed, make your dreams come true. But learn to curate your creations. Not every single output is a masterpiece.

[Quicksand-Lover-7]

Is posting 3-5 variations of a concept fine, or is it just 1 image per concept.

[dlodoski]

Is posting 3-5 variations of a concept fine, or is it just 1 image per concept.

You posted 20. It's best not to be obtuse right now.

"20 Images Of Hawaiian Hula Girls Sinking In Quicksand.
Post by Quicksand-Lover-7 » Sat May 10, 2025 7:15 pm"

Five is fine. Three is better. One is best. Nuff said.