Has AI stagnated producers and creativity

[Mr. X]

Right now the market is inundated by producers who make picture sets and little movies using AI. The problem is most of the pic sets are variations on the same poses which one gets their fill of pretty quickly. Same with the movies. NOTHING tells a story. No interesting camera angles or set pieces. All are like someone in a photo studio snapping in front of a camera.

but we're flooded with these and no one seems to be really making original art. Even the DLSITE and pixiv scenes are inundated with ai sets with mange and hentai falling away. No stories, no real plots.

We seem to be in this dry period in which we have actually lost creative producers. No one makes anything anymore.

[yc-batgirl]

About a year ago, I tried to use some AI software since everything is going that way. Testing them and starting to learn them.

Like you said, poses and camera angles are not great. What I hate the most is emotions, like only two are available, smiling and sad. I have to say lightning is not great, a front view of a model in an outdoor scene gets the same lightning when looking around. The sun is following the camera.

Maybe producers are trying to learn AI to save big money: models ($$$), sets, lights, cameras...

I'm still using my old 3D software to create pictures and even if my pictures don't look as real as what AI can do, I'm proud of what I do.
They come from my imagination and creativity. The easy way is not always the best way

[yc-batgirl]

If I may say something else, don't forget that we are living in a world where everything is self-centered.

Too many people are even cheating to gets the most ''clicks'', ''likes'' and views. Some of best cheaters are using bots at trying to become famous.

[Visitor]

Go on Deviant Art ad see people dump those hundred images that are almost the same. Even when someone is creative enough to have them change enough to tell a story you see alternating mistakes like costume whole, costume torn, and then back again. Or Wonder Woman nude but her lasso is hanging off her beltless waist where it would normally be.

[five_red]

There are two problems:

• The tools themselves have historically not been sophisticated enough to process complex prompts for non-trivial scenes.
• The barrier to entry for creating content has been lowered, resulting in a lot of lazy people churning out slop that just feeds their particular kink, without attempting to be original or imaginative.

So Google just made their Imagen 4 model freely available, with minimal or no usage limits, via ImageFX. (Yes, it has censorship, before anyone asks.) The model is significantly better at processing complex requests, even quirky stuff that has only limited references for from images online. Case in point: as a test I asked it to generate an invisible character, a hollow suit of clothes posed as if being worn by a person. It did this straight off without any problem, creating a flawless photo-realistic image. I then made the scene more complex by getting the invisible character to do various actions (walk, open a door, wrestle with another character) and it was able to generate these scenes with only the occasional hiccup on some images. What's impressive about this is that there won't have been too many images it could have scrapped online of invisible characters to feed the model, and yet it was able to create good results consistently with different poses and clothing choices. After experimenting I've discovered Imagen 4 is also able to better handle specific instructions for ripped or damaged clothes (She-Hulk type characters), interactions between characters (looking at someone or something, pointing at things, fighting/struggling/wrestling, touching/holding/hugging, etc.), scenery and props (ajar doors, objects falling, characters handling objects in a particular way, telephones and cars being period specific to 1940s/50s/60s/etc.), and a range of other improvements.

So the models are starting to get better at interpreting prompts. The problem is to unlock this sophistication requires a lot of specific details be put into the prompt. Specific clothes, poses, lighting, scenery, camera positions, etc. etc. And the vast majority of users are just too lazy to write all that detail. Some of my own prompts can run to a dozen or so sentences, many of which will be run-on sentences with multiple clauses. The vast majority of prompts I've seen listed on DeviantArt is just two or three relatively short sentences. So no wonder we keep getting bland, unimaginative, slop.

And it also has to be mentioned that, while the AI can act as a poor-man's artist, it cannot replace the author. The reason I write long prompts is because I want to create little stories with each picture. I don't want pic after pic of static lifeless portraits -- I want funny little vignettes. (Of course, it helps that I've also read about comicbook lettering, and know how to create fairly faithful speech balloons and effects.) So I put the time in, and I experiment to see what the limits of each model are. Very very few users, judging by the prompts listed on sites like DeviantArt, bother. And -- truth be told -- very few users will bother in future, even as the models get better and better. The models are already at a point where they are capable of far more than just the bland portraits that fill up web sites today, but the majority of users are too lazy and/or too unimaginative to want to put the effort in to write complex detailed prompts.

Just my 2c.

R5

[Femina]

Is this REALLY surprising to anyone?

I feel like people have been warning that this was coming ever since AI was first announced?

Of COURSE it's stagnating art. Who the f'k wants to work their asses off to produce Art and present it to the world only for it to disappear in a sea of someone's lazy AI generations?

This is just the logical outcome of automation?

[Mr. X]

Its funny how we 3D people were accused of the same laziness. Granted I will always have respect for someone who draws by hand. Now GOOD AI art is difficult. Many passes. Knowing the tools etc. But same was for 3D. Its not just posing a figure then spinning the camera around.

[Mr. X]

Another area that is getting flooded is AI voice products. They have like 3 pics but all voice acting, usually in Japanese. Places get flooded with these.

[Opaque]

The SHiP genre is dying off as the old timers either had their fill or are dying off themselves. The consumers who remains have the genre all figured out. When they got in decades ago almost anything was a buy. Now it's "I'm not buying unless this, that and the other". It's difficult for a creative producer who doesn't have the funds to just do it for fun.
Low barrier to entry to AI creation means any no talent hack can just churn out that stuff day in and day out. They don't care about consistency or quality control. If it's close enough it's close enough. "Free" to produce and free to upload. Slap on a 5 dollar price tag and if you get 10 people to buy some of that garbage then hey 50 bucks!
That said it's not like the SHiP genre isn't filled with slop made by and with real people. Personally with the AI stuff, I'm seeing myself going that route and "produce" my own stuff through AI tools. Simply because the genre has largely left me behind in what it caters to and with AI tools that are tuned properly with LoRAs and all that jazz, you can make everything to your liking.
Just go try Grok Imagine image to video for example(it's free) and upload an image and try to puppeteer a scenario using the image. Sure it's 6 seconds of 480p and censored for NSFW(still useable for classic SHiP tropes), but you will get an idea of what is ahead of us when it comes to AI images and video.

[CIA]

*

[Void]

I don't know if the SHiP genre is dying off... I mean, maybe what we would consider classic SHiP is dying off, but I think the rough interest in powerful heroines getting into trouble is probably an ever-green one, and super heroines are always going to be a strong expression of that.

That said, there's no hiding from the fact that the scene is a pale, withering shadow of what it was ten years ago - when this forum was FAR for active, and when new content was being made at an astounding rate from an astounding number of producers, to a really engaged and active crowd of consumers. People aging out is probably a big factor, but the squeezing of profit margins, the extra competition, and the insane ease of piracy will all of played a big role.

AI is kind of a mixed bag, in terms of whether it's helpful in this case. For my purposes, it's allowed me to generate insanely good cover art for heroine peril books virtually for free, where before I would have had to spend significant money to commission one-off images. I mean, I know this is killing artists, and I feel awful for them, and I'm part of the problem by turning from them to this, but this plays a major role in enabling me to be a one-man band publishing company where my costs are very low and I can still cobble together something with the optics of a proper book - and which nails the image I want (after a lot of iteration and tweaking). Without AI generated art, it would be far less appealing to attempt to write books for release. Also, copilot, and probably most big AI platforms, is a solid 8/10 proof-reader that works instantly and for free. In this sense, I find AI to be an incredible resource that expands creative horizons rather than shrinks them.

[Femina]

Its funny how we 3D people were accused of the same laziness. Granted I will always have respect for someone who draws by hand. Now GOOD AI art is difficult. Many passes. Knowing the tools etc. But same was for 3D. Its not just posing a figure then spinning the camera around.

It really isn't the same at all? EVERY element of 3D art is artwork. The end-'user' of the assets may not themselves be responsible for the entire piece to be sure, but human beings are involved in every step. Someone programmed every line of code in the program. Someone modeled all the 3D assets for use. Someone took those assets and modified their textures/geometry/etc. as desired, someone framed the scene, set the lighting, oriented the assets in precisely the desired poses. Who all was involved, how many of those steps were accomplished personally vs collaboratively is all up in the air. Some people do nearly ALL of it, modeling, texturing, posing etc. and thus make it ENTIRELY a solo artistic effort (aside from perhaps having not devolved the program themselves.

At no point in any of this, is the artistic process left up to a random algorithm process. When the image is finished 'rendering' the user looks it over, sees what they want tweaked... and DIRECTLY acts to tweak it not unlike erasing digital brush strokes and changing them, or looking over the canvas and deducing its limitations and what one can personally do with it.

Nowhere would I qualify MOST 3D artists on par with hand drawn artists or sculptors, songwriters etc. But comparing 3D artists with AI artists skips the step where all AI artists are playing a 'Gacha' game with the algorithm every time they get a result. NEVER is the result exactly what you intended, never.

There is no algorithm in 3D art, the outcome is the expressed INTENT of the artists involved from the bottom up with NO inanimate decision making at any point.

AI is GREAT candy. I in no way mean to say it's not. I like candy, I've eaten the candy... but it's not art, and you aren't an artist when the machine makes a picture for you... and it's 100% stagnating artistic creativity. You want to be an artist. Go draw a nifty hopscotch course on your driveway, you'll be more of an artist than you ever will just shitting out AI candy.

Some pretexts... One might use AI to INSPIRE them towards producing ACTUAL art of their own. In which case it might be considered an asset for an artist. But nothing the AI has a direct hand in is art.

[Dazzle1]

AI has stagnated all facets of business

[Mr. X]

AI has stagnated all facets of business

I do use Claude AI for coding. Does great for things like cleaning up project files or making a web page or analyzing code to see what section does what.

[HiDef]

Its funny how we 3D people were accused of the same laziness. Granted I will always have respect for someone who draws by hand. Now GOOD AI art is difficult. Many passes. Knowing the tools etc. But same was for 3D. Its not just posing a figure then spinning the camera around.

I have to disagree with you, they're two different animals.

With AI generators all one has to do is properly word the prompt to avoid censorship and out pops a character, then add that character to an image to video engine and out pops an animation. Having to reword prompts or continuing to generate until you get something usable, I don't equate to artistry.

I've been doing photomanips and animations for many years with many different tools as they became available. Getting rid of Windows for MacOS was a big change for the better. I equate Manips to being Dr. Frankenstein, sometimes having to create missing limbs or cut them from multiple sources. There are a lot brush techniques I've learned over the years that makes things easier but it's generally a lot more involved than typing prompts. I don't draw much but I can paint.

When I first started animating it was like claymation, it's painstakingly laborious. I had to first use a graphic editor and a lot of different techniques to simulate movement. I had to create individual frames with slight alterations and then compile them into a video editor timeline which is a different type of skill.

Creating and syncing audio tracks to animations can still be a PIA so I usually opt for no audio.

To address the topic, I'll touch on three issues.

1.Censorship is certainly playing a big role in stifling creativity in both AI and live-action. Clips4sale has been reduced to little more than wedgies and foot smelling.

2. From the AI perspective, it takes a lot of VRAM and processor power to produce anything more than short clips which are then compiled into a video editor. Even short clips can consume a thousand watts or more of power for the duration. Now multiply that by how many passes it takes to get something usable and you get a sky-high electric bill.

3. AI is still in the beta stage when it comes to video creation.

Oh and one more thing, future hardware and software innovations will certainly be a driving force behind AI video creation.

[Dazzle1]

AI has stagnated all facets of business

I do use Claude AI for coding. Does great for things like cleaning up project files or making a web page or analyzing code to see what section does what.

But does it help for story ideas?

[Mr. X]

AI has stagnated all facets of business

I do use Claude AI for coding. Does great for things like cleaning up project files or making a web page or analyzing code to see what section does what.

But does it help for story ideas?

Never tried it for that. Just pointing out an example of AI use. I'm not banging rocks together

[villainofsooperhero]

Not sure if this will work but I'll try it.
I'm guilty of about all of what has been mentioned above. I'm a long time photographer / producer of this stuff and I still shoot and create live action movies but the AI stuff has been fun. For me, it's a revelation taking old photos that I may have shot 20 years ago, that might be 640x480 and turning them into 1080p movies. But holy shit does it piss people off. The artists hate it, obviously. The members of the community hate it, because it comes off as cheating and "not real". But for me, it's a lot of fun. I agree, there's a lot of frustrating content being put up because of the 5 second renders and repetitive stuff. But this is a sample of one I did last week using one of my shoots. It's kind of cool. (It's on my server so it might take some time to load at over 7 minutes...but it's definitely not a 5 second clip.
[media]https://superheroinesdemise.com/video/the_heist.mp4[/media]
So, using the photo shoot I have the images which followed a story line and I can prompt what I want in terms of movement, and then riff off that. It's not text to video or text to image so it gives me more continuity with the character and the costume since I'm feeding ai a real photograph. With all that said, it's not going to replace real video tomorrow. Maybe eventually but not yet. As I get ready to shoot with a new model next week I'm reminded of one thing I like about ai, no tattoos. Lotta tattoos these days on models and that's not very supergirl like. ha ha.

[HiDef]

Not sure if this will work but I'll try it.
I'm guilty of about all of what has been mentioned above. I'm a long time photographer / producer of this stuff and I still shoot and create live action movies but the AI stuff has been fun. For me, it's a revelation taking old photos that I may have shot 20 years ago, that might be 640x480 and turning them into 1080p movies. But holy shit does it piss people off. The artists hate it, obviously. The members of the community hate it, because it comes off as cheating and "not real". But for me, it's a lot of fun. I agree, there's a lot of frustrating content being put up because of the 5 second renders and repetitive stuff. But this is a sample of one I did last week using one of my shoots. It's kind of cool. (It's on my server so it might take some time to load at over 7 minutes...but it's definitely not a 5 second clip.
[media]https://superheroinesdemise.com/video/the_heist.mp4[/media]
So, using the photo shoot I have the images which followed a story line and I can prompt what I want in terms of movement, and then riff off that. It's not text to video or text to image so it gives me more continuity with the character and the costume since I'm feeding ai a real photograph. With all that said, it's not going to replace real video tomorrow. Maybe eventually but not yet. As I get ready to shoot with a new model next week I'm reminded of one thing I like about ai, no tattoos. Lotta tattoos these days on models and that's not very supergirl like. ha ha.

This is right up my ally. I love bringing new life to old SHIP, upscaling is kinda my thing. I've done a lot of it but I keep a lot of it to myself so not to "piss people off" especially the original producer. I'm also a big fan of image to video so I can control the continuity of the characters and build a better storyline of my own creation, not AI generated. Wren vs the MilkMaster 2000 could be a lot of fun if it were not for the oppressive censorship filters.
I agree 100% on the tattoo issue. I don't like tattoos on women and I absolutely hate them on superheroines.

[lionbadger]

We seem to be in this dry period in which we have actually lost creative producers. No one makes anything anymore.

He says, effectively pining for the mide 2010s when there were 14 30 minute films a week of.... "wonder woman being chloroformed and debooted!"

on a moe serious note, the twin evils of inflation and conservativism have stangnated a lot of things or rolled back the open spaces we used to have. Everything is more expensive and a lot of people are just finding out that 'loan interest' is a thing again, knocks on to everything. I think we might actually be lucky that AI is there to fill the gap that would otherwise be just a gaping wasteland of the same wonder woman and supergirl pieces over and over again.

[Opaque]

As I get ready to shoot with a new model next week I'm reminded of one thing I like about ai, no tattoos. Lotta tattoos these days on models and that's not very supergirl like. ha ha.

Yeah, it's most unfortunate that it's very hard to get a woman without tattoos these days. At least that's what a lot of producers have been saying for years and I can see it in their films.
Tattoos can be removed using AI on images and the results will be very usuable without looking altered. I don't think it's there yet with video. I'm sure there will come a time when tattoos can easily be removed from video.

[Visitor]

Removing stuff in editing is done in major studio productions, but they have the money and ability to hire enough people to do the work. After all they need to do it for wire work and other scenes where you need to substitute CGI for an another object. But getting it down in cost and power for this isn't there.

[plymoutha]

You can check on website: Crazy Cattle 3D

[RedMountain]

My guess is rising costs etc. have probably stagnated producers more than anything else, that and the crackdown/forced registry for certain sites where videos are purchased in some states. I've seen a few producers mention that it's just not worth it to put in all the costs to make a video when hardly anyone is going to buy it or the consumers are getting so picky these days about the content they want. Obviously you can bypass the age verification stuff with VPN, but I know a lot of people who simply don't even bother anymore now that they can no longer go to certain sites that want you to submit a photo ID or selfie photo, etc. to verify your age.

[Visitor]

Google recently added age verification and that applies to YouTube. I'm not sure how secure their storage of verifying information is when two of the choices are driver's license type and credit card.

[Abductorenmadrid]

I've looked at AI for sure and for some applications regarding art and it's worked for what I want. And I mean that in a Dungeons and Dragons DM viewpoint, where I wanted to be able to give my players a prop. Need an image of some NPC to put on a wanted poster? Done. Need mock artwork for an inn's sign above the entrance? Done ... Need some prop sketch notes with designs for some arcane device ? Done.

For all that, it's been great. It wasn't something I would commission an artist to create and wasn't anything I might be able to pull off the net. So, yeah, AI works for that.

For my SHiP stuff, I've looked, and not really gotten anything to work for me. Perhaps I'm not using the right tools, I don't know. What I do sense is a lack of control over output. Not spot-on-specific, at least. Output might come out right, or technically right, as the prompt indicated, but, it's never specifically what was in my head. Only true artists can create that.

I have similar issues with my own manipulation work. I can alter, bend and distort things to fit to what I want, but the core of things will only be what I can scavenge. I can 'create' something quite adjacent to what an original image had - make new realities in that space - sometimes they can look quite good - but I can't make anything from scratch. I seriously think though, I have more control than what some AI users have over their own creations. Sure, they may get something photoreal, even move like video, but is it exactly what they wanted?

No, I think real artists, real actors, real producers can get their visions as they want, we others are just passengers trying to make what we can, work.

[hagarb]

Today we have a lot of content created by AI, and the vast majority is either not great or consists of 300 thousand small variations of the same things. There is indeed some very good AI-generated content, but it ends up getting lost in the middle of all this. And this ease of generating images and videos discourages many people from trying the traditional path of filming or manual drawing. The result is that it's much harder to find something good, and what good content is found doesn't come close to the best human productions that used to pop up here and there. The creation tools are improving a lot, but I understand that there is a lack of curation solutions to find good quality content with less effort.

[MightyHypnotic]

Most of the time if I see something has been created with AI I just skip it. I really liked it in the beginning when AI was new and I was curious what it could do. It's only going to get better but for now, I prefer live action videos/photos with real models.

I've seen a few standouts on DA who are producing some interesting AI videos. One of my favorites is
BlunderBoy1959
https://www.deviantart.com/blunderboy1959

I'm not into the subject matter, but the execution is really good. I wouldn't mind trying to make a Blue Angel or Knightwoman movie using that tech.

[Malice]

Personally, I think AI opens new horizons for creators.

It allows to produce visuals that would have required multiple specialized skills before.

It doesn’t replace creativity, and there is a human behind the prompt.

[Mr. X]

I think right now it just can't get you there. In 6 months to a year that will be different. I suspect someone could create a story board app in which you describe each panel and have character descriptions and then it does the comic for you.

But right now we're mostly seeing just the same pic with 100 different variations put out as products and some people are just flooding the market with these.

This is pretty cool.
https://www.deviantart.com/mauler77/art ... 1259186977

[MightyHypnotic]

I think right now it just can't get you there. In 6 months to a year that will be different. I suspect someone could create a story board app in which you describe each panel and have character descriptions and then it does the comic for you.

But right now we're mostly seeing just the same pic with 100 different variations put out as products and some people are just flooding the market with these.

This is pretty cool.
https://www.deviantart.com/mauler77/art ... 1259186977

Yeah, that shows a lot of promise.

[SHL]

Right now the market is inundated by producers who make picture sets and little movies using AI. The problem is most of the pic sets are variations on the same poses which one gets their fill of pretty quickly. Same with the movies. NOTHING tells a story. No interesting camera angles or set pieces. All are like someone in a photo studio snapping in front of a camera.

but we're flooded with these and no one seems to be really making original art. Even the DLSITE and pixiv scenes are inundated with ai sets with mange and hentai falling away. No stories, no real plots.

We seem to be in this dry period in which we have actually lost creative producers. No one makes anything anymore.

Most people who use AI to make images have never in their life developed being a creative, cause they show all their work. Every artist I have ever known don't show off their practice. Posting 100 variations of the same image is a quick way to kill the dopamine in your audience.

We have no young blood when it comes to SHIP filmmaking. And there is no way for any new creators to even try. The cost of producing films is too high and only getting higher (and more risky) and the amount of people who can afford to buy films is on a downward trend, especially in Trump's 'grab em by the pussy' economy.

I don't know I see the point in complaining about it though. If my production budgets have to be cut down in cost - I can't really be more creative in storytelling. If I don't have as much money as I use to when it comes to what I could put into a production - having a bunch of people criticize me for not being as creative isn't going to magically make me have more money to spend.

The more volatile the economy becomes - the less consumer spending there is for films - more content producers will exit the genre from lack of sales - the ones who stay will be making choices to guarantee results (panic creation) and that usually means the death of creativity.

I predict it's only going to get worse. So, grab onto your butts

[Mr. X]

Most people who use AI to make images have never in their life developed being a creative, cause they show all their work. Every artist I have ever known don't show off their practice. Posting 100 variations of the same image is a quick way to kill the dopamine in your audience.

We have no young blood when it comes to SHIP filmmaking. And there is no way for any new creators to even try. The cost of producing films is too high and only getting higher (and more risky) and the amount of people who can afford to buy films is on a downward trend, especially in Trump's 'grab em by the pussy' economy.

I don't know I see the point in complaining about it though. If my production budgets have to be cut down in cost - I can't really be more creative in storytelling. If I don't have as much money as I use to when it comes to what I could put into a production - having a bunch of people criticize me for not being as creative isn't going to magically make me have more money to spend.

The more volatile the economy becomes - the less consumer spending there is for films - more content producers will exit the genre from lack of sales - the ones who stay will be making choices to guarantee results (panic creation) and that usually means the death of creativity.

I predict it's only going to get worse. So, grab onto your butts

Yeah I liken this stage similar to the Lara Croft, square but stage of video games at this moment. Also Only Fans led to a lack of models. Why shoot for someone when you can do it for yourself and you get money without elaborate scripts and costumes.

[superbia19872]

Right now the market is inundated by producers who make picture sets and little movies using AI. The problem is most of the pic sets are variations on the same poses which one gets their fill of pretty quickly. Same with the movies. NOTHING tells a story. No interesting camera angles or set pieces. All are like someone in a photo studio snapping in front of a camera.

but we're flooded with these and no one seems to be really making original art. Even the DLSITE and pixiv scenes are inundated with ai sets with mange and hentai falling away. No stories, no real plots.

We seem to be in this dry period in which we have actually lost creative producers. No one makes anything anymore.

Honestly, it was all going downhill before AI ever hit the scene for me. This is largely do to the hard turn into rape fantasies, which if you can indulge them responsibly, that's you. No judgement. But i was starved of the content that I want and bluestone was kind of the last holdout, but they have spiraled in quality in my opinion. The writing never impressed me. The dialogue was always too gratuitous for how poorly written and delivered it was, and their habit of telling not showing has been immersion breaking for a long time, but the content still had the tension to excite me.

I definitely agree with SHL that most AI "creators" have no idea what they're doing, but it's weird that they can't tell that their content is largely unusable. From what I can tell based on the months I've spent developing my process, they don't direct the acting, give the models zero context, and pack entirely too many beats into a single prompt so the model rushes and generally overacts everything. I've encountered 2 or 3 who absolutely have stronger processes than I do, but good god, I wish everyone else would be more discerning and honest with themselves. They just aren't taking the time to iterate enough or make adjustments to get it there, which isn't completely unfair because it IS expensive right now, but it's still cheaper than ordering customs. I think people are also dead set on having clean visuals, and i really believe that right now you have to choose between good performances with occasional glitches and clean visuals. For me that is not worse than the state of the genre when I found it in 2007 when the clips were wooden and moody anyway.

I've wasted so much time giving them chances that I'm just over it and that's a bummer because I certainly can't afford to produce much content and it takes me about a month to generate the shots and edit them for 10-15 minute videos. It should be unleashing creativity but instant gratification is really stifling the vast majority.

Anyway, no hard feelings for my support in the community dissolving. I just never find anything that I want anymore. I know several of you are still creative and competent.