Too Much AI, Not Enough Creativity

[HiDef]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

[sam84]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers. Nulls Brawl APK Download

Is anyone else tired of all the repetitive AI content?

thanks for share good information

[sam84]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

thanks for share amazing blog

[SHL]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

Ironically it appears a bot responded to you twice, so... the end is nigh

I personally find almost nothing interesting about this AI nonsense thats flooding this community, but I don't often talk about it publicly cause I figure those who do it might have a thin skin about it

I would be surprised if people actually loved it

[thegreengiant]

They really should really have a toggle that removes any ai. Google images too.

[hagarb]

Yeah, the repetitive thing is very annoying. I really like AI, I have excellent results using it, but I only post what I like the most to avoid flooding the community with all the minor variations around the same context.

[Mr. X]

One account won artist of the year for making $$ by flooding with AI art and by shear volume ended up making more money. Can't imagine the wasted drive space.

Also this is happening on OFs where producers are churning out AI videos. In fact one company sells a turnkey system on how to make a channel and the art.

This seems like the house of cards for OF girls may collapse and this time next year may see a massive reduction of OF girls that are replaced by AI. The irony is their content will be used to generate the AI content.

I think the winners will be the people who can combine posing with 3D and an AI pass with face replacement. Use the 3D as a skeleton so the AI has something to work with.

[hpibd]

Wow you Inspire me Amen God Bless him

Kidding, I'm just a boring old human

Seriously though, I personally think AI has some serious potential and can be really useful, but not to overrun the art community with low effort imagery.

[Mr. X]

Seriously though, I personally think AI has some serious potential and can be really useful, but not to overrun the art community with low effort imagery.

That's the issue. Yeah the flood of nonsense. And right now you can't really do a story with it, just variations of the same pic. Its the flooding that's the issue.

[HiDef]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

Ironically it appears a bot responded to you twice, so... the end is nigh

I personally find almost nothing interesting about this AI nonsense thats flooding this community, but I don't often talk about it publicly cause I figure those who do it might have a thin skin about it

I would be surprised if people actually loved it

Same here.

I have no issues with using AI for improving the quality of original live action content but the strictly AI generated stuff is a waste of server space, IMO. Yeah, I'm sure there are AI fanboys who love it but it's not for me.

[HiDef]

One account won artist of the year for making $$ by flooding with AI art and by shear volume ended up making more money. Can't imagine the wasted drive space.

Also this is happening on OFs where producers are churning out AI videos. In fact one company sells a turnkey system on how to make a channel and the art.

This seems like the house of cards for OF girls may collapse and this time next year may see a massive reduction of OF girls that are replaced by AI. The irony is their content will be used to generate the AI content.

I think the winners will be the people who can combine posing with 3D and an AI pass with face replacement. Use the 3D as a skeleton so the AI has something to work with.

That's a slap in the face to real artists.

Where do you draw the line for OF copyright infringement? That's always in the back of my mind when doing manips. I'm careful to only use content that has been freely distributed. Although when creating photo stories, I may steal a few body parts from paid content.

[MightyHypnotic]

AI generated content has completely overtaken the superheroine genre over at DeviantArt. I think I can safely say that over 95% of it is AI generated. There's no artistry or talent involved, they just have a chat with an AI bot and out pops several options for them to choose from.

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

thanks for share good information

I guess the answer to "Are you a robot" was yes. Maybe that safeguard allows entry based on honesty?

Are you a robot?
Why, yes, yes I am a robot.
Well then, by all means, enter.

[xxxwarriorxxx]

I block every single one that I come across. It filters them out my feed quickly. I also do not care for AI in this format.

[lionbadger]

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

No, because before it was hundreds of people churning out the same wonder woman, super girl and if lucky the odd batgirl image. The community has some of the worst fans for "show me the same thing itis" so nothing has really changed and its not like you can't sit down with a pen and draw your own stuff

[Mr. X]

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

No, because before it was hundreds of people churning out the same wonder woman, super girl and if lucky the odd batgirl image. The community has some of the worst fans for "show me the same thing itis" so nothing has really changed and its not like you can't sit down with a pen and draw your own stuff

I think the problem there is the lack of ability to do any kind of constructive negative feedback to deter this stuff. Plus free is a pretty good price so people tolerate a lot.

[HiDef]

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

No, because before it was hundreds of people churning out the same wonder woman, super girl and if lucky the odd batgirl image. The community has some of the worst fans for "show me the same thing itis" so nothing has really changed and its not like you can't sit down with a pen and draw your own stuff

The point I'm trying to make is, I don't like the flooding of forums with repetitive AI images that requires no creativity or talent whatsoever. That's why they sell hard drives so people can upgrade their own machines. I have to assume that limits will soon be imposed due to servers being overwhelmed? What kind of a idiot posts nearly 20.000 images in 4 months time?

[Mr. X]

Just 1 example of the disproportionate amount of AI content is a Deviant who joined only 4 months ago but already has nearly 20,000 AI generated images uploaded. That's an average of 167 posts per day, every day for 4 months. And there are many more examples of this excessive and redundant flooding of the servers.

Is anyone else tired of all the repetitive AI content?

No, because before it was hundreds of people churning out the same wonder woman, super girl and if lucky the odd batgirl image. The community has some of the worst fans for "show me the same thing itis" so nothing has really changed and its not like you can't sit down with a pen and draw your own stuff

The point I'm trying to make is, I don't like the flooding of forums with repetitive AI images that requires no creativity or talent whatsoever. That's why they sell hard drives so people can upgrade their own machines. I have to assume that limits will soon be imposed due to servers being overwhelmed? What kind of a idiot posts nearly 20.000 images in 4 months time?

Yeah Renderosity had to put a limit on AI art in its galleries. Disc space isn't infinite. Also pixiv has limits.

On a side note one thing I hate about Deviant Art's pay thing is you have no idea what is in the item. Its just a blurred thumbnail and, unless the artist posts samples, you don't know. Why would anyone pay $5 for something they don't know what's in it.

[HiDef]

Now there are some real talented artists doing animated 3D content. Sort of what Mr.X mentioned above.

Although I prefer real meat as opposed to Beyond Burgers, I don't mind eating vegan once in a while but not everyday for breakfast, lunch and dinner.

Here's my favorite 3D animation so far: https://rule34.xxx/index.php?page=post& ... id=7239888 All credit to original creator.

[HiDef]

Yeah Renderosity had to put a limit on AI art in its galleries. Disc space isn't infinite. Also pixiv has limits.

On a side note one thing I hate about Deviant Art's pay thing is you have no idea what is in the item. Its just a blurred thumbnail and, unless the artist posts samples, you don't know. Why would anyone pay $5 for something they don't know what's in it.

100% agree, no preview means no buy for me. I also want to know resolution and quality before buying.

[TheWatcher]

Being someone who posts AI images on DeviantArt, I completely agree! I was one of the first to start posting AI images, of mostly Supergirl in peril images, on DA... under the name Necromancer-Rev, but before that I was writing stories on here and I have been a fan of the superheroine in peril genre for many many years... dating back to the hay days of SuperheroineCentral and The Wizards Lair. I've thought about charging for my content, but the Wizard always kept his free, so I've done the same. Having said that, the number of AI generated superheroine in peril content is ridiculous... you have to scroll through so much shit "art" where... for Supergirl, the colors of her costume aren't right and its basically hundreds of the same fucking bullshit images with the same AI image prompt. I have over 20k images that I've generated but I've only posted around 2500 of them over years because 1. I don't want to flood people's feed and 2. not all of the images I generated are that good... not even all the ones I've posted are all that good. Then there are online friends of mine that were posting great AI generated content with very complex prompts who got banned from DA for some bullshit reasons which has definitely altered what gets posted on DA. The bottom line is, there is way too much content being posted on DA that is basically hundreds of the same images with slight differences. DA recently implemented a bundle of images option where on the paid side you can bundle 50 images into one post and on the unpaid side you can bundle 10 into one post... I've been trying to use this feature more and more to limit how much I block up the feed, but I wish more people did the same. Posting that way also limits how many people see your content because they have to scroll through a ton of shit images to get to the content they want to see. It also impacts actual artists who are drawing or creating digital content that isn't AI driven and that definitely fucking sucks. I'm not sure what the cure is but hopefully it gets better and more sorted out. As note and if any of yall are interested, here's my gallery of mostly Supergirl in peril images... hopefully, it makes you think that not all AI image posters are bad but if not, I still wish you well! https://www.deviantart.com/necromancer-rev/gallery

[HiDef]

Being someone who posts AI images on DeviantArt, I completely agree! I was one of the first to start posting AI images, of mostly Supergirl in peril images, on DA... under the name Necromancer-Rev, but before that I was writing stories on here and I have been a fan of the superheroine in peril genre for many many years... dating back to the hay days of SuperheroineCentral and The Wizards Lair. I've thought about charging for my content, but the Wizard always kept his free, so I've done the same. Having said that, the number of AI generated superheroine in peril content is ridiculous... you have to scroll through so much shit "art" where... for Supergirl, the colors of her costume aren't right and its basically hundreds of the same fucking bullshit images with the same AI image prompt. I have over 20k images that I've generated but I've only posted around 2500 of them over years because 1. I don't want to flood people's feed and 2. not all of the images I generated are that good... not even all the ones I've posted are all that good. Then there are online friends of mine that were posting great AI generated content with very complex prompts who got banned from DA for some bullshit reasons which has definitely altered what gets posted on DA. The bottom line is, there is way too much content being posted on DA that is basically hundreds of the same images with slight differences. DA recently implemented a bundle of images option where on the paid side you can bundle 50 images into one post and on the unpaid side you can bundle 10 into one post... I've been trying to use this feature more and more to limit how much I block up the feed, but I wish more people did the same. Posting that way also limits how many people see your content because they have to scroll through a ton of shit images to get to the content they want to see. It also impacts actual artists who are drawing or creating digital content that isn't AI driven and that definitely fucking sucks. I'm not sure what the cure is but hopefully it gets better and more sorted out. As note and if any of yall are interested, here's my gallery of mostly Supergirl in peril images... hopefully, it makes you think that not all AI image posters are bad but if not, I still wish you well! https://www.deviantart.com/necromancer-rev/gallery

When did they implement bundling for free accounts? I have not been able to bundle images into a single post. When I try to do so, I'm prompted to upgrade to Core.

[personnal]

For me the tools are not efficient enough but they will be. I already have some very good results and I think people like what I do. And it is not so easy , because before having a good picture it is a lot and a lot of tries , corrections, modifications and more. Here is one of my last picture I like . Hope you will like

[personnal]

For me the tools are not efficient enough but they will be. I already have some very good results and I think people like what I do. And it is not so easy , because before having a good picture it is a lot and a lot of tries , corrections, modifications and more. Here is one of my last picture I like . Hope you will like

[McGheeny]

I have to admit I have been intrigued and yes even dabbled in the AI. But I can say that I create my own art, whether with Poser or Gimp, it is something I have been doing for years. Lately I have been working with AI to enhance my images, not create them. I use AI as just one more tool to give the viewer the best product I can produce. Many times I still go back to Gimp with the AI enhancement to fix details AI misses or errors it creates. It's a fun tool but in the end I prefer my creations over computer generated images that don't always come out the way you describe them. But of course, that is just my take on AI.

[BlackOut]

It's an interesting topic, but the way I see it, it's just another tool in the tool box.

[sneakly]

Bad art on the Internet? No! It can't be so....

Block the people posting 500 a day. It's a tool and I think it will mature as people get better at it and the software is refined. A year ago hands looked tentacles 85% of the time and a quarter of the people had three or more legs. Eventually talented people will figure out how to use it effectively. It was the same when poser and all the other rendering tools came out. The people starting out have to figure out how to use the tools somewhere and get feed back so they can become better. How many stories are you not reading on DA after looking at the first sentence?

[five_red]

When did they implement bundling for free accounts? I have not been able to bundle images into a single post. When I try to do so, I'm prompted to upgrade to Core.

They turned it on a few weeks back.

Submit you first image. When you are taken to the submit screen where you can fill out the post's meta data, add additional images on the right hand side. Either click the PLUS button under the image preview, or simply drag/drop images from your computer into the submit screen.


R5

[Mr. X]

When did they implement bundling for free accounts? I have not been able to bundle images into a single post. When I try to do so, I'm prompted to upgrade to Core.

They turned it on a few weeks back.

Submit you first image. When you are taken to the submit screen where you can fill out the post's meta data, add additional images on the right hand side. Either click the PLUS button under the image preview, or simply drag/drop images from your computer into the submit screen.


R5

I think you can also just drag and drop multiple pics at once

[HiDef]

When did they implement bundling for free accounts? I have not been able to bundle images into a single post. When I try to do so, I'm prompted to upgrade to Core.

They turned it on a few weeks back.

Submit you first image. When you are taken to the submit screen where you can fill out the post's meta data, add additional images on the right hand side. Either click the PLUS button under the image preview, or simply drag/drop images from your computer into the submit screen.


R5

I think you can also just drag and drop multiple pics at once

I guess I should clarify what I mean? I've been able to upload multiple files in batches for a while now but I have to publish/submit them one at a time. When publishing photo stories, most times I could just use a preset for all the images and publish them in batches but I can't do that unless I upgrade to a Core membership.

[Mr. X]

I guess I should clarify what I mean? I've been able to upload multiple files in batches for a while now but I have to publish/submit them one at a time. When publishing photo stories, most times I could just use a preset for all the images and publish them in batches but I can't do that unless I upgrade to a Core membership.

Yeah I remember you could save things as a default. I don't post that much multiple pics so I don't deal with this.

[Bladebur]

AI is a tool. You can prompt it and generate 160 daily images that are mostly variations of the same portrait. With a suitable LoRA, those can even be some kind of bondage or peril position full of mistakes that doesn't make a lot of sense. But you can also take your time, learn the tool, and the results will be way better. AI mistakes can be fixed or iterated, important stuff like faces or hands can be rebuilt at high resolution to fix them, and you can compose multiple AI images and objects using photobashing techniques.

I used to think DA was being destroyed by the AI flood, but now I'm not so sure. I've been taking more effort lately to create better pictures and I'm getting similar or better exposition compared to what I was getting before the AI when I did only 3D stuff. That being said, it's certainly discouraging when you get a favorite and then you check the follower's favorites only to find your image buried under 100 similar AI portraits.

I don't know about DA's revenue streams, perhaps those have been hurt by this. In any case, they never made any sense to me (selling an image without being able to see it? Or 'selling' an image without giving any usage/copy rights at all to the new 'owner'?). I suspect Patreon/Subscribestar are way better for this kind of thing, with DA used as a promotion site just to find customers.

Batgirl tight rope.png (3.23 MiB) Viewed 109738 times

[HiDef]

AI is a tool. You can prompt it and generate 160 daily images that are mostly variations of the same portrait. With a suitable LoRA, those can even be some kind of bondage or peril position full of mistakes that doesn't make a lot of sense. But you can also take your time, learn the tool, and the results will be way better. AI mistakes can be fixed or iterated, important stuff like faces or hands can be rebuilt at high resolution to fix them, and you can compose multiple AI images and objects using photobashing techniques.

I used to think DA was being destroyed by the AI flood, but now I'm not so sure. I've been taking more effort lately to create better pictures and I'm getting similar or better exposition compared to what I was getting before the AI when I did only 3D stuff. That being said, it's certainly discouraging when you get a favorite and then you check the follower's favorites only to find your image buried under 100 similar AI portraits.

I don't know about DA's revenue streams, perhaps those have been hurt by this. In any case, they never made any sense to me (selling an image without being able to see it? Or 'selling' an image without giving any usage/copy rights at all to the new 'owner'?). I suspect Patreon/Subscribestar are way better for this kind of thing, with DA used as a promotion site just to find customers.


Batgirl tight rope.png

Yes, a tool for artists. Not for idiots who flood forums with images they did not create themselves. Sorry, but manipulating an AI bot to produce images is not art.

As for likes, favorites, etc., what's with all the incessant thanking of others for liking content? Let me get this straight, I create something which someone else likes or adds to their favorites, now I'm expected to thank them for thanking me? I don't create anything for others unless I'm paid to do so. Displaying my art on a web-forum is showcasing that which would otherwise only be seen by myself. It's not an ego-trip for me. Now I'm not saying that those "likes" mean nothing, they do incentivize me to continue adding content.

I originally opened an account over at DA, 9 years ago to host manips I did for a cosplayer friend of mine who lives in a foreign country where it is illegal to possess, produce, distribute or even view pornographic imagery. I didn't do it for anyone else or for likes.

My latest entries were to showcase imagery for extra income, but I decided against charging money for manips due to copyright infringement. Now, I may begin sharing original content I created myself to sell on Patreon or some other pay-site in the future? And yes, I may use AI to improve quality but not for generating content.

[Bladebur]

Yes, a tool for artists. Not for idiots who flood forums with images they did not create themselves. Sorry, but manipulating an AI bot to produce images is not art.

As for likes, favorites, etc., what's with all the incessant thanking of others for liking content? Let me get this straight, I create something which someone else likes or adds to their favorites, now I'm expected to thank them for thanking me? I don't create anything for others unless I'm paid to do so. Displaying my art on a web-forum is showcasing that which would otherwise only be seen by myself. It's not an ego-trip for me. Now I'm not saying that those "likes" mean nothing, they do incentivize me to continue adding content.

I originally opened an account over at DA, 9 years ago to host manips I did for a cosplayer friend of mine who lives in a foreign country where it is illegal to possess, produce, distribute or even view pornographic imagery. I didn't do it for anyone else or for likes.

My latest entries were to showcase imagery for extra income, but I decided against charging money for manips due to copyright infringement. Now, I may begin sharing original content I created myself to sell on Patreon or some other pay-site in the future? And yes, I may use AI to improve quality but not for generating content.

Wow, you certainly have an axe to grind.

In my opinion, the 'X is not art' declaration is a bit misplaced, it's something I've heard many times before... Photomanipulation is not art, it's just mixing or crudely coloring existing images made by others. Poser/DAZ is not art, it's just mixing 3D models and props made by others. Photoshopping is not art, it's just altering existing photos made by others. Digital painting with references is not art, it's just copying bits from existing images made by others. Painting in a canvas from live models or landscapes is not art, it's just copying what you see in reality.

At the end of the day, any form of art you can imagine has been criticized at some point for not being art: it takes too little effort, it's just not original, it conveys no emotion, it's soulless, it is too functional, etc.

Personally, I couldn't care less. If it's not art, then ok, I don't do art. I make titillating images of superheroines in compromised situations for my own enjoyment and share them for the enjoyment of others. By the way, I like getting feedback from people with similar tastes and I think thanking them is nice. There is a lot of that ego-stroking social hamster wheel thing going on, which may end up being a terrible thing in the grand scheme of things (I think social networks are a huge net negative for humanity at large) but meh.

I've considered sometimes to sell art, but I just don't see the numbers. It would be (for me) way too much effort to make something others could be somewhat interested in, for too little return. Skill issue, probably.

[HiDef]

Yes, a tool for artists. Not for idiots who flood forums with images they did not create themselves. Sorry, but manipulating an AI bot to produce images is not art.

As for likes, favorites, etc., what's with all the incessant thanking of others for liking content? Let me get this straight, I create something which someone else likes or adds to their favorites, now I'm expected to thank them for thanking me? I don't create anything for others unless I'm paid to do so. Displaying my art on a web-forum is showcasing that which would otherwise only be seen by myself. It's not an ego-trip for me. Now I'm not saying that those "likes" mean nothing, they do incentivize me to continue adding content.

I originally opened an account over at DA, 9 years ago to host manips I did for a cosplayer friend of mine who lives in a foreign country where it is illegal to possess, produce, distribute or even view pornographic imagery. I didn't do it for anyone else or for likes.

My latest entries were to showcase imagery for extra income, but I decided against charging money for manips due to copyright infringement. Now, I may begin sharing original content I created myself to sell on Patreon or some other pay-site in the future? And yes, I may use AI to improve quality but not for generating content.

Wow, you certainly have an axe to grind.

In my opinion, the 'X is not art' declaration is a bit misplaced, it's something I've heard many times before... Photomanipulation is not art, it's just mixing or crudely coloring existing images made by others. Poser/DAZ is not art, it's just mixing 3D models and props made by others. Photoshopping is not art, it's just altering existing photos made by others. Digital painting with references is not art, it's just copying bits from existing images made by others. Painting in a canvas from live models or landscapes is not art, it's just copying what you see in reality.

At the end of the day, any form of art you can imagine has been criticized at some point for not being art: it takes too little effort, it's just not original, it conveys no emotion, it's soulless, it is too functional, etc.

Personally, I couldn't care less. If it's not art, then ok, I don't do art. I make titillating images of superheroines in compromised situations for my own enjoyment and share them for the enjoyment of others. By the way, I like getting feedback from people with similar tastes and I think thanking them is nice. There is a lot of that ego-stroking social hamster wheel thing going on, which may end up being a terrible thing in the grand scheme of things (I think social networks are a huge net negative for humanity at large) but meh.

I've considered sometimes to sell art, but I just don't see the numbers. It would be (for me) way too much effort to make something others could be somewhat interested in, for too little return. Skill issue, probably.

"axe to grind", no not exactly, although it may seem that way to someone who posts AI content. My last response was more of a jab back at the 4 times in this thread someone stated the obvious, reminding me that AI is a "tool" while completely ignoring the OP. I didn't start this thread to bash AI and if you read my original post, you'll find it's about flooding forums with repetitive AI generated content.

BTW, I have since used DA's content filter to block AI content.

[BlackOut]

I was meaning to expand on my blatantly simple statement. As someone who uses AI frequently, I most certainly believe that it takes some measure of talent to create the art that I (and several others) share on DA (or other platforms). I'm sure some will disagree with that sentiment, but as someone who has a degree in the arts and has spent countless hours sketching, painting, sculpting, and doodling away during my younger years I feel pretty secure in my opinion.

But granted, that's just my opinion, and others aren't going to agree. Just the same way people didn't agree that a lot of other technological breakthroughs weren't really art when they filtered their way into the mainstream. That being said, as to the OP's point, yeah, AI has taken over the vast majority of Deviant Art - and personally while it doesn't really bother me, there is the option to block it (as you stated you had done).

I do agree that it is kind of ridiculous to see some of the accounts out there just flooding their galleries with mountains of images that I don't think they even take a moment to look over and critique. I find it pretty easy to ignore though. I follow the people who create stuff that I like, and my opinion is probably skewed since my focus is on animation (which is a much smaller niche in the AI art scene on Deviant Art)

[Femina]

You aren't an artist unless you're directly responsible for the decision making process of the art. Art should say something about the artist (or artists plural if it's a collaborative piece). Something of the artist themselves should be in the art... be that something as simple and direct as your brushstrokes... or as complicated as evidence that a medieval sculptor had a left arm injury based upon the chisel work in a sculpture... BUT If it's actually IMPOSSIBLE to link the artist to the art in some fundamental way that can't be made clear to an observer who can see or extrapolate those links SPECIFICALLY with the artist... than it isn't their art.

The AI is the ENTIRE decision process of the art in an LLM, at best you're an artistic 'consultant' on the output. There's a mild argument to insist that there is always a middle 'medium' in art... the pencil, the chisel, the mouse and paint programs, 3D modeling/animation program etc. etc... but none of those have any actual power or control over the decision making process that goes into the artistic piece, they are simply the tools with which an artist uses to deliver their vision or intent into the art... in AI imagery as it currently exists, the AI actually has TOTAL control over the output, and the human being may only *suggest* what they would like it to create, nothing of themselves exists in or on the output unless they've created and trained the LLM entirely with their own artwork they have manually created, which is the ONLY scenario at all in which I can find any leeway whatsoever that an LLM is producing any form of output that can even circumstantially be described as 'your' art.

There is possibly a world in the future where some sort of AI tools CAN create real art, and the people who use them artists, what that process or tools could be I don't know... but they aren't LLMs as they currently exist. LLM's directly take the brushstrokes of others, and directly translate them into pixels as the LLM sees fit with loose guidance (VERY LOOSE in some cases) from a human being suggesting conditions (conditions the LLM quite often ignores, only further proving the human being has no actual control over the output beyond what the LLM allows them to have).

Wanna be a REAL artist? It's not that hard. Go outside and draw a hopscotch game on your sidewalk. Draw a stick figure animation on a notepad. Carve a crude cross out of a hunk of bark... Even if it looks awful, it's Yours! YOU made it, congratulations you're an artist. You'll become more of an artist drawing stick figure's in a flipbook than you ever will be asking the AI to create any number of beautiful or thought provoking Picasso-esque images FOR you.

It's a fun distraction, It's a decent entertainment technology to goof off with to kill an afternoon. It is NOT currently an artists toolkit...... and it's absolutely, objectively, flooding the internet with GARBAGE, which is evidence of actual harm the technology is inflicting... enough to suggest to me that responsible use of AI image technology ought be entirely for personal enjoyment... and NEVER shared beyond the occasional e-mail or DM to a friend, while filling up spaces intended for true art with them is rapidly proving to be an irresponsible and damaging practice.

IMO it takes a considerable amount of arrogance to label oneself an artist for writing up a prompt in a textbox then claiming credit for the image the algorithm shits out, a considerable amount more to place oneself on equal footing with the creative process and workload of true artists for doing so, and EVEN MORE so beyond that to charge money for the images that the AI created for you........... but I mean, fuck, people gotta eat and if someone's stupid enough to buy a pack of ten images your computer created for you they could just go and get the computer to make for them... and it feeds you for the week who am I to judge? Just don't call yourself an artist for it... CON artist maybe.

[Bladebur]

The AI is the ENTIRE decision process of the art in an LLM, at best you're an artistic 'consultant' on the output. There's a mild argument to insist that there is always a middle 'medium' in art... the pencil, the chisel, the mouse and paint programs, 3D modeling/animation program etc. etc... but none of those have any actual power or control over the decision making process that goes into the artistic piece, they are simply the tools with which an artist uses to deliver their vision or intent into the art... in AI imagery as it currently exists, the AI actually has TOTAL control over the output, and the human being may only *suggest* what they would like it to create, nothing of themselves exists in or on the output unless they've created and trained the LLM entirely with their own artwork they have manually created, which is the ONLY scenario at all in which I can find any leeway whatsoever that an LLM is producing any form of output that can even circumstantially be described as 'your' art.

Well, perhaps that's the case in the cheapest of AI productions, but it is not necessarily true.

Consider Poser images, which are often made from existing models, clothes, props, and even existing pose and lighting assets. I don't see a huge difference between choosing those assets or choosing prompt tags. With Poser, the best content IMHO is that where Poser is just one part of the workflow: clothes are customized, new textures are painted, new props are modelled for the image, and the render is brought to Photoshop where it is augmented by manual edits, color corrections, etc. It's the same thing with AI. There is an 'asset search' phase (finding proper LORAs and models, prompting, generating hundreds of test images), a 'custom asset building' phase (training your own LORAs, preparing control poses, drawing sketches or 3D renders for depth control) and an 'edit' phase (bring the AI results to Photoshop for composition and manual editing) that is arguably much heavier compared to the Poser equivalent because there is always a ton of AI BS to clean up.

A lot of people are flooding the Internet with cheap AI where they do nothing of this. This is sad and diminishes the potential of the tool. The same thing happened with Poser.

[Abductorenmadrid]

Regarding AI sourced images in DA - yeah it's an epic problem and apart from limiting post frequency I don't know what else they could do to stop it. Maybe (gasp - use AI to weed out the AI content) - or how about sweep it into a new "Deviant AIrt.com"

I would be a bit more welcoming if it seemed like there was at least a little work from the "prompter", make the AI output part of something bigger, but I've not seen that much. Christ, if someone was capable of coding the automation of a process they could get a bot to make a prompt, grab the output and upload it and the prompter wouldnt need to do much at all!

<Superheroine_name>in <peril_type_> with <villain_name> at <scene_setting> + <photoreal_please>...

With all the AI content around no wonder it's hard to get much thanks or feedback for anything these days!

AEM

[Femina]

The AI is the ENTIRE decision process of the art in an LLM, at best you're an artistic 'consultant' on the output. There's a mild argument to insist that there is always a middle 'medium' in art... the pencil, the chisel, the mouse and paint programs, 3D modeling/animation program etc. etc... but none of those have any actual power or control over the decision making process that goes into the artistic piece, they are simply the tools with which an artist uses to deliver their vision or intent into the art... in AI imagery as it currently exists, the AI actually has TOTAL control over the output, and the human being may only *suggest* what they would like it to create, nothing of themselves exists in or on the output unless they've created and trained the LLM entirely with their own artwork they have manually created, which is the ONLY scenario at all in which I can find any leeway whatsoever that an LLM is producing any form of output that can even circumstantially be described as 'your' art.

Well, perhaps that's the case in the cheapest of AI productions, but it is not necessarily true.

Consider Poser images, which are often made from existing models, clothes, props, and even existing pose and lighting assets. I don't see a huge difference between choosing those assets or choosing prompt tags. With Poser, the best content IMHO is that where Poser is just one part of the workflow: clothes are customized, new textures are painted, new props are modelled for the image, and the render is brought to Photoshop where it is augmented by manual edits, color corrections, etc. It's the same thing with AI. There is an 'asset search' phase (finding proper LORAs and models, prompting, generating hundreds of test images), a 'custom asset building' phase (training your own LORAs, preparing control poses, drawing sketches or 3D renders for depth control) and an 'edit' phase (bring the AI results to Photoshop for composition and manual editing) that is arguably much heavier compared to the Poser equivalent because there is always a ton of AI BS to clean up.

A lot of people are flooding the Internet with cheap AI where they do nothing of this. This is sad and diminishes the potential of the tool. The same thing happened with Poser.

Don't misunderstand me. I'm not here preaching that AI will NEVER be used to create art, I'm not one of those art snobs stuffing their nose in the air calling computers out as 'cheating' in special effects thus denying TRON the Oscar for REVOLUTIONARY FX.

I'm expressing that LLM image models CANNOT create art and make you an artist. When you use Poser models, you have measures of control over the outcome. Sure, you may have to use premade models and articles that other's have made... but you don't even HAVE to do that. You c an model and rig YOUR OWN models and props. You have control over the angles of the scene, the lighting, you can strip all the surfaces of their textures and carefully cultivate new ones (your OWN typically to find the look you're going for), Poser/Daz is typically akin to a 'Collaborative Art Project' You are not typically responsible for the ENTIRETY of the output artwork involved in the piece, but you have a DIRECT hand in the artistic decisions, you are responsible for SOME of the artistic output. The amount of your own personal contributions in this sense vary from person to person in accordance with their skills... but even if it's simply choosing camera angles and tweaking the poses into the exact specifications you want them to be, you've contributed DIRECT input into the output.

LLMs have NO Direct human involvement at all outside of the coding gone into the program. Gradually we are seeing AI programs that aim to address this with more and more additional tools, and I am in NO WAY against this. AI models experimenting with pose controls to FORCE the AI to abide by certain parameters... but the current scope of LLM's as they are still retain the ability to veto all these parameters and spit out a garbage noisy, scrambled, mess of spaghetti images that could NEVER happen if you had direct personal control and input into the piece. When Daz render's an image I decide isn't good enough, it's usually cause I notice little things like... a toe isn't posed quite how I envisioned it, or the lighting is bouncing off a surface I don't want... and I can directly get in there, tweak it, and the problem is solved, rinse and repeat until the output is EXACTLY what I want. When an LLM does this, ALL you can do is hit 'generate' again and hope that whatever it shits out next is PERFECT, because it will NOT address the tiny imperfections that caught your eye that brought you to 'generate' again, it will simply output something entirely new.

There is also a psychological element of art that LLM users don't factor here. Art isn't JUST about 'physical' workload. Art is often HARD, which is why most people aren't just always doing art. Artists go into ruts where they lack inspiration. Writer's get writer's block etc... Element's of psychological hang-ups that aren't associable with an AI algorithm that does the heavy lifting of the most taxing element of artistic expression FOR you. At no point is one so mentally exhausted by the LLM's generative process that they find themselves incapable of clicking 'Generate' again. That aspect of artistic endeavor doesn't exist in an LLM process because it isn't art.

I'm not arguing not to use it. I'm not arguing not to seek and test AI tools in the search for the ones that will be useful in producing ACTUAL art... I'm simply arguing that the current trend on the world wide web of shitting out 5000 AI images from ones personal musings into the internet is becoming an OBJECTIVELY damaging practice... and that it's a damaging practice being undertaken from a faulty reasoning standard by folks who want to believe that doing so makes them artists. It doesn't, and every day is proving that entertaining the idea that they are is actively harmful to the state of our current hubs of artistic expression.

I'm not saying don't AI. I'm saying AI responsibly... and to please please PLEASE not irresponsibly shatter Deviantart or wherever else you frequent with sixty AI images an afternoon of AI garbage. Keep your AI stuff to yourself, ENTERTAIN yourself with it... but don't ruin art spaces with it. Let the technology breathe, let it evolve, and one day, eventually, it will produce tools that DO make art, but the LLM cannot do this. The LLM is no more than a monitor showing us pretty colors that spur our endorphins, and wont ever be anything more... but with luck and a little elbow grease MAYBE what comes after it WILL......... but only if we don't allow LLMs to destroy our ability to produce and deliver art entirely before that time comes to pass. Once Deviantart becomes nothing more than a server filled to the brim with 99% AI generated images... there won't be any 'fixing' it, and no one else is going to pick up Deviantart's sword to create a NEW space just to watch IT destroyed by AI slop as well... we need to find a way to get a handle on this, and that starts with AI tech bros recognizing that they need to use their tools more responsibly.

[five_red]

I've written about this elsewhere. The fact that AI can churn out lots of images cheaply doesn't means it has no value, but the quality of what it pumps out depends upon the who is doing the prompting.

I've used the example of Watchmen, drawn by Dave Gibbons, written by Alan Moore. Gibbons was drawing panels based on Moore's script. In essence Moore (the writer) was providing 'prompts' for Gibbons (the artist) to follow, in the same way that AI users provide prompts for the AI to draw. So is Moore no better than an AI user, leeching off the skill of Gibbons? The key difference here is that Moore wasn't just asking for simple portraits of characters, he was providing a scenario for those characters that told a story. It takes zero skill to ask for a flat portrait of a character -- the skill comes in story telling, context, nuance.

064_018_image_fx_final.jpg (385.13 KiB) Viewed 109598 times

Because of AI, for the first time in history, a lot of people without serious artistic skills are able to generate pretty high quality images. And thrilled with their new found abilities they are spewing the results all over sites like Deviantart. It is kind of like that phenomena when people who have recently become parents for the first time feel the need to share every moment of their parenting journey with friends and family. Their excitement causes them to massively overshare. And I suspect the same is currently happening with AI art -- the thrill and novelty of being able to create images without any artistic skill (coupled with the fact that the results keep getting better and better with each month) is causing them to massively overshare.

I suspect this phenomena will reduce once the novelty has worn off -- but we'll see.

063_311_image_fx_final.jpg (321.72 KiB) Viewed 109598 times

What we really need is not just a reduction in the number of spammy images, but an increase in their dramatic quality. AI users need to understand that AI doesn't turn them into Dave Gibbons, it turns them into Alan Moore. The skill with AI is creating an image that tells a story. Anybody can prompt for flat uninspired portraits. The true art is in using lighting, camera position, character posing, and a whole load of other factors, to make the image dynamic what gives a sense of something happening.

_99207319-7c7d-4256-a9d1-a155e98d3283_final.jpg (339.63 KiB) Viewed 109598 times

_4611e0ef-ff82-4397-b43d-bdd646ba4eac_final.jpg (280.84 KiB) Viewed 109598 times

_53385000-d9af-4a03-9de2-eb75c31edc8e_final.jpg (267.26 KiB) Viewed 109598 times

R5

[Get Your Exxon]

It’s been a long time since I’ve posted.

Part of the reason is exactly what’s been touched upon by this thread. Despite all the endless content too much of what’s out there lacks intention or creative vision.

It’s funny because I remember how hard it used to be to find this type of content. In fact, as a Supergirl fan, there was a time when every morsel felt precious. And yet, despite all the promise and possibility of today’s tech, I too find that it’s missing intentionality.

Especially in a genre that often boils down to giving a pretty girl, with amazing abilities, a bad day, a certain reinvention is necessary. Maybe I’m wrong, but I appreciate when an artist attempts something new. And despite all the glitz and glamour of today’s tech it feels harder, not easier, to find good stuff. In many ways, I see more creativity in the past. I’m not sure if that’s because of the limitation of the tech or because I’m not a teenager anymore.

Please don’t misunderstand, I hunger for new material. For example, I’m really enjoying what Dr. Evil is doing with Dark Reign. For those who are unfamiliar he’s absolutely found a great way to reinvent a familiar beat(ing). And while I may not agree with every decision or outcome, there is no questioning the skill or creativity. Actually, I wish this forum would applaud it more because it’s honestly some of the best art (and writing) out there.

After all, so many of the great creators that helped build this genre are gone. I’m not sure why. Hopefully, it’s just life taking them on a better path. But, yeah, good work is hard to find. And, again, thank you for posting. It’s definitely a topic worth discussing.

[HiDef]

Regarding AI sourced images in DA - yeah it's an epic problem and apart from limiting post frequency I don't know what else they could do to stop it. Maybe (gasp - use AI to weed out the AI content) - or how about sweep it into a new "Deviant AIrt.com"

I would be a bit more welcoming if it seemed like there was at least a little work from the "prompter", make the AI output part of something bigger, but I've not seen that much. Christ, if someone was capable of coding the automation of a process they could get a bot to make a prompt, grab the output and upload it and the prompter wouldnt need to do much at all!

<Superheroine_name>in <peril_type_> with <villain_name> at <scene_setting> + <photoreal_please>...

With all the AI content around no wonder it's hard to get much thanks or feedback for anything these days!

AEM

Yeah, that's it! That's what it is, "AIrt". I like it. There's a joke in there somewhere?

Oops, I just sh-AIrt-ed and now my DA page is splattered with crappy images.

[Mr. X]

I do know the good AI does take lots of work. Running enhancement Loras, doing post process etc. Its not just a prompt. Saying its just a prompt is the same criticism I get doing renders "Just toss up a pose and render.. how hard is that"

The problem is the flood of stuff just from prompts in which the user just fired off 100 renders and didn't even remove the extra fingers etc and then posting that, that is the problem.

Same in the 3D days. Posting tons of poor renders doesn't help with acceptance.

Even pre-computer quality was always favored over quantity. We all have that favorite pic in which there is just one version.

[SHL]

Does no one understand how hard it is for me to make a good prompt? I slave away all day thinking about what keys to type... maybe... maybe ill write the word cleave gag... if only... I could find the courage...

I do think it's funny that multiple people posted their AI images in this conversation as a means to show off or prove a point. Read the fucking room LOL

Oh... thats a good prompt actually... Create an image, inspired by 1950s comic books... a superheroine wears... a cleave gag... and she can't read the room... oh nooooo.... don't bully me I am going to cummmmmmmm!

[five_red]

I do think it's funny that multiple people posted their AI images in this conversation as a means to show off or prove a point. Read the fucking room LOL

You didn't actually read the thread, did you?

(A lot of the commentary is about how prompting is easy, but what is lacking is creative vision. And that's why we're getting a lot of lazy, repetitive, uninspired images. So -- assuming I've understood your rather haphazard way of expressing yourself -- the point attempting to be proven by the people in this thread, in part agrees with you.)


R5

[SHL]

You didn't actually read the thread, did you?

(A lot of the commentary is about how prompting is easy, but what is lacking is creative vision. And that's why we're getting a lot of lazy, repetitive, uninspired images. So -- assuming I've understood your rather haphazard way of expressing yourself -- the point attempting to be proven by the people in this thread, in part agrees with you.)

I did, but to quote Carly Simon 'You're so vain, I bet you think this song is about you, don't you?

[HiDef]

I do think it's funny that multiple people posted their AI images in this conversation as a means to show off or prove a point. Read the fucking room LOL

You didn't actually read the thread, did you?

(A lot of the commentary is about how prompting is easy, but what is lacking is creative vision. And that's why we're getting a lot of lazy, repetitive, uninspired images. So -- assuming I've understood your rather haphazard way of expressing yourself -- the point attempting to be proven by the people in this thread, in part agrees with you.)


R5

Lack of creative vision on the developers part, yes. But from an artist's viewpoint AI has hit a wall. I don't see it improving unless the developers allow more input from the enduser. They need to improve upon the user interface (UI) of AI or lack thereof.

[five_red]

I did, but to quote Carly Simon 'You're so vain, I bet you think this song is about you, don't you?

You didn't reference me in what you wrote. And I didn't reference myself in my reply (except to make a pragmatic comment about whether I'd understood your disjointed comments.) So... perhaps you're seeing things that just ain't there..?

What you previously wrote was somewhat incoherent, but seemed to suggest that you didn't consider merely typing words into an AI made for quality images. In fact a lot of discussion here seems to agree with that sentiment, at least in part (if indeed that was the point you were making -- your argument was a tad ambiguous.) There does seem to be a general consensus here that too many people are generating lots of low-effort images, rather than taking time to experiment by refining and enhancing their prompts, then picking the one or two results that work. Few people here seem to be dead-set against AI in an absolute sense, only the flooding of low-effort images which is currently plaguing some sites.

A lot of people are writing prompts that are only one short sentence long, often around a dozen words. Ideally, to get an image that conveys any kind of emotion or dynamism, they should be writing something closer to a paragraph at least. Then revising and extending the prompt text to correct and nudge the AI in particular directions, before doing clean-up by getting the AI to redraw parts of a good candidate image, and post-production work in Photoshop to tidy everything up. But that's not what is happening all too often. Too many people are feeding in short prompts, and spamming the output with little filtering straight out onto Deviantart, etc. They're just generating multiples of flat portraits, rather than trying to tell a story with a single well crafted image.

R5

[five_red]

Lack of creative vision on the developers part, yes. But from an artist's viewpoint AI has hit a wall. I don't see it improving unless the developers allow more input from the enduser. They need to improve upon the user interface (UI) of AI or lack thereof.

I agree to some extent. Tools are currently far too hit-n-miss. You roll a dice and hope something good comes up, but currently you have little precise control over how to fix problems or make specific changes to existing images. But I also partially disagree too. With a bit of effort, and some lateral thinking, you can coax AI to get more consistent results. Not perfect -- but better! Sometimes it is about getting creative with how you word prompts (using synonyms and reordering the clauses.) Other times you need to learn to use a specific tool more productively (use the thumbs up/down buttons in ImageFX to tell the AI when it is getting warmer or colder.) You also need to experiment with levels of detail.

The problem is I don't think many people want to spend the time to experiment. Which I understand, because it can sometimes feel like herding cats. I just wish they were a bit more discerning about what they shared on DA.


R5

[SHL]

I did, but to quote Carly Simon 'You're so vain, I bet you think this song is about you, don't you?

You didn't reference me in what you wrote. And I didn't reference myself in my reply (except to make a pragmatic comment about whether I'd understood your disjointed comments.) So... perhaps you're seeing things that just ain't there..?

What you previously wrote was somewhat incoherent, but seemed to suggest that you didn't consider merely typing words into an AI made for quality images. In fact a lot of discussion here seems to agree with that sentiment, at least in part (if indeed that was the point you were making -- your argument was a tad ambiguous.) There does seem to be a general consensus here that too many people are generating lots of low-effort images, rather than taking time to experiment by refining and enhancing their prompts, then picking the one or two results that work. Few people here seem to be dead-set against AI in an absolute sense, only the flooding of low-effort images which is currently plaguing some sites.

A lot of people are writing prompts that are only one short sentence long, often around a dozen words. Ideally, to get an image that conveys any kind of emotion or dynamism, they should be writing something closer to a paragraph at least. Then revising and extending the prompt text to correct and nudge the AI in particular directions, before doing clean-up by getting the AI to redraw parts of a good candidate image, and post-production work in Photoshop to tidy everything up. But that's not what is happening all too often. Too many people are feeding in short prompts, and spamming the output with little filtering straight out onto Deviantart, etc. They're just generating multiples of flat portraits, rather than trying to tell a story with a single well crafted image.

R5

I am not reading all of that, save it for a prompt