so, background, I’m currently workshopping a Disco Elysium/Touhou Project fusion fanfic (Disco concept, Gensokyo setting) and, since I already pay for NovelAI for recreational purposes, I figured I’d generate some character/skill portraits for it to accentuate the presentation of it.
Here's some of the better ones as of yet, alongside the one in the thumbnail.
However, I am conflicted on actually using them.
I figured I’d outsource the question because I’ve been thinking about it for a week and am no closer to resolving the internal contradiction. Here’s what I’ve been thinking thus far:
Concerns: I have moral concerns about the scraping and the ‘ask-for-forgiveness-not-permission’ philosophy that is central to these major generative models. I also think, in the context of it’s use within the hands of the bourgeoisie, AI is used as a tool of displacement and discipline. I also know that there is generally a negative perception of people who use AI image generation, for mostly valid reasons as most self-proclaimed ‘AI artists’ are coomers/techbros who make it a mission to discredit and disparage human artists. Even though I don’t fall into this category, I’m worried that I’d open up a vector of harassment where there otherwise wouldn’t be one if I include them, and I don’t want to deal with that.
Benefits: It adds to the vibes of the fic, and I think there’s a tangible enhancement with adding visuals to it. I also like what I’ve output so far.
Caveats: I don’t plan to make money off of this. I don’t have the disposable income to commission portraits for what ultimately is a fanwork that I don’t plan to profit off of and don’t have enough investment in to sink money into. The concerns are specifically what are holding me back in adding them, because it would be nice to add a visual component.
I just wanted to get a vibe check on here for this use case. Want to see if it will help me resolve the internal contradiction I’ve tangled myself in 
edit: I’ll err on the side of caution and drop the visual aspect of it, thanks for the feedback
The battle is already mostly lost. Capitalism will subsume all of it.
This is no different than the low skilled jobs in the manufacturing sector of industrialized countries being exported and replaced by cheap labor from the developing world. The “locally made” and “quality crafted” products being replaced by cheap (but no less functional) products being churned out by the assembly line in some third world country.
This is also no different to what happened in the developing world when ag giants flooded their countries with GMO seeds and livestocks and displaced the locally cultivated produces. The local farmers were simply unable to out-compete with the superior yield of the genetically modified products, and were poised to lose out if they don’t also give in to multi-national agricultural businesses.
And throughout such transformative changes over the last few decades under neoliberalism, the tech sector and the art/creative industries had been able to stay relatively unscathed. But they, too, will not be able to avert the same fate that befell the other industries. Handcrafted, creative arts being displaced by machine (silicon) generated pieces without soul.
Without protectionist strategy in place, all of these local businesses will be devoured and subsumed by the Free Market.
So, in a sense, it is not wrong to say that there is no ethical consumption under capitalism (you are already doing it with food and consumer goods), so long as you are not explicitly supporting the capitalist exploiter. But ultimately the solution can only come from radical and systemic changes to the entire system itself.
I don’t plan to make money off of this
Then there’s nothing to feel bad about IMO.
People should feel bad using something that requires stolen art to function
Generative AI is a labor issue, not a property issue, except as it pertains to the models themselves which as they seek to enclose all of human culture should themselves be public domain and owned by all.The easiest way to prove this is this simple thought experiment: if a major corporation, say Disney, could have a model produced exclusively with material it owned or otherwise licensed for the explicit purpose of model training, would that it make it ok for that model to then be used in their animation studios or licensed out to other studios? And the answer is a resounding no: it doesn’t matter that they own the property involved, because the harm the model is doing is to skilled labor. So clearly the question of property is just a diversion that keeps the discourse within boundaries that are favorable to huge corporations who can easily own all they need.
That’s why my stance has been and remains that generative AI should be a copyright poison pill: the models should be public domain even if trained on licensed art, the direct outputs of them should be public domain, works that include them should be public domain including the parts that aren’t AI generated, and as an extra penalty all the IP within a work containing generative AI should become public domain as well. So if, say, Disney were to use some generative AI in a Star Wars movie, not only should that movie become public domain but the entire Star Wars IP should as well, along with all the Disney trademarks they toss in in splash screens at the front just because fuck them.
I think that’s a fundamental misunderstanding of how generative models work and what stealing is
👍
Nah I get it, even if you’re not making money off of it I understand feeling weird about using bootleg versions of other people’s creativity.
I don’t see any issue. You are not making money off of the product, nor did you intend to pay an artist in the first place.
My only concern is that it contributes towards the normalization of generative AI use, but that is a battle already lost.
and anyone using the normalization argument had better have the purest media history mankind has ever seen, or else they’re doing a heckin normalization too!!
I don’t understand what you mean by this comment. Could you reword it in clearer terms?
Sure - If the argument is that using and engaging with generative AI normalizes its use, pretty much everyone consuming media at the moment is normalizing negative practices in much the same way by consuming it.
That being said, I think it’s okay to draw lines in the sand for causes more personally relevant to you, but you have to acknowledge that you’re doing that on a personal whim and make peace with other people’s whims in that sense. Otherwise, you’re either an ascetic or you’re a hypocrite.
I kinda think of it like programmer art, I want to replace it with something better eventually but having anything there helps keep the project flowing, which is great for motivation. If the project gets far enough that it’s worthy of help from other disciplines then I 100% wouldn’t leave it in there, but for quick/small/personal projects who really gives a shit?
Tohou ppl are generally pretty nice. I wouldn’t worry about it.
If it really bugs you treat em like developer graphics and try to chase down someone with an appropriate style who will do your character portraits for real.
The stuff you’ve got already would be a fantastic base for an artist to work from, so even if it’s just used to communicate more clearly it’s not a waste to use ai.
E: also don’t listen to anyone itt slagging you for using the weird amoral new tool. Jesus Christ motherfuckers straight up can’t tell the difference between a comrades labor of love and the broader corporate effort to literally reshape the truth before our eyes.
Smdh my dick head.
The stuff you’ve got already would be a fantastic base for an artist to work from, so even if it’s just used to communicate more clearly it’s not a waste to use ai.
i see we’ve gotten to the clean room design stage
You’ll definitely alienate at least some people if you use AI art, and there’s a good chance you’ll be harassed at some point if you post an AI art project outside of an AI friendly space. I do think that a lot of people don’t have a innate or built-up aversion, so I doubt it would be universally a negative.
Personally, I think your art direction with the AI imagery is nice, and some people will appreciate it. A lot of AI projects have effectively no art direction and just fall to the mean (which contributes to the aversion to commercial AI media products), so you’ve already cleared one barrier. Since it’s not interactive media, I think you don’t necessarily need to have a visual element, but I see how it could help you stand out.
we’re so backMildly negative. You’re not taking work away from an artist, because there was never any chance you were going to pay one. But it’s still kinda “eh”.
I can’t form a coherent link right now, but that makes me think about some of the arguments in favor of piracy.
All AI art is shitty to use regardless of why anyone is using it
I’d never willingly download, share, recommend, or speak positively about something I knew involved AI-generated images. My stance on this sort of thing is always just get an artist to create the art you want. You’re already paying for this shitty stolen AI-generated garbage, pay an artist.
You’re already paying for this shitty stolen AI-generated garbage, pay an artist.
acting like the cost of a cloud based ai image generator is anywhere near equivalent to commissioning a dozen+ portraits from an artist is very disingenuous
It does cost more to get an artist to make art for you than to get an algorithm to steal art for you, yes. That isn’t a very good argument for it being ok.
when the argument is “if you’re already paying for x, you should just pay for y instead”, x being much less expensive than y is a very important factor in that argument
Much less expensive but also morally wrong
It costs hundreds of times more. “It does cost more” is the understatement of the century.
My point is that paying any amount of money at all for AI-generated theft is bad. I don’t care if they’re paying you, it still sucks.
Idk what about Fiver? Although it’s probably full of “AI” “artists” now.
If you do use the generated portraits, be sure to put that you used generated imagery in the description on whatever you end up hosting the game on.
Several games have used AI art without disclosing it and everyone found out eventually prompting them to release patches that remove it all.
That being said, theres definitely nothing wrong with using it during development of you’re working on a solo or small scale project. As having a whole concept department is pretty expensive or time consuming and if concept art isn’t your strong suit it’s difficult to do.
Ai “art”
I already pay for NovelAI for recreational purposes
That money could be spent on a human artist to make the art. They could also make it look consistent, whereas the AI stuff you show has an inconsistent style. Some stuff for example has distinct “line-art” whereas some relies entirely on color and shading to indicate contours and delineations, although on your end you could use the difference selectively to set a mood and make it seem deliberate (like line-art style normally, switching to color-and-shadow for emotionally heavy scenes, particular characters that carry a special significance, etc.).
Also there was a post here a while back about the energy cost of generating these AI images, it’s apparently an average of “1 smartphone charge” per go: https://hexbear.net/post/1274630
I don’t know how that compares to the energy used by an artist making a bunch of stuff on their computer over several hours / days, so that might not be much of a point, but it’s something to consider.
Also, I’m not the only person who just won’t fuck with a project that they know to have AI-generated content. Keep that in mind as far as how much / what kind of audience you want your fan-project to have.
11.49 kWh per 1000 images worst-case
keep in mind that 11 Wh is basically a rounding error, as far as amounts of power go. Equivalent to 40 seconds of a microwave, 8 minutes of a burning tealight, or 8 cents on my power bill ($0.07/kWh). Likely similar order of magnitude to digital artist using a lower-power device for hours. Worth considering when thinking about large policy but not in assessing individual actions.
sorry i’m unable to resist plugging these things into Wolfram
edit: hey I don’t like this upvote ratio. This isn’t meant to say that OP should use AI art or even as a substantial response to GnastyGnuts, per-image energy is the least important part of this convo
OP already said that there is no budget to pay an artist to draw all the portraits. The choice isn’t between generative AI or human artist, the choice is between generative AI or nothing.
NovelAI starts at $10 per month, while a human artist would charge hundreds of dollars for all the portraits. There is no budget for that, so it was never an option in the first place.
Maybe it’s just me being a boomer millennial, but I wouldn’t be able to tell any of these are made by AI if you just show them to me individually. I can only tell when I see them all at once since the art styles are simultaneously too different and also kinda not distinctive enough. All of these are well within good-enough range for some bougie fuck to crank out cereal box and toothpaste design using AI art.
As machine learning gets more sophisticated, the amount of time spend to create good-enough art will drop since the average time spend is largely (average time spend to generate prompt)/(percentage of non-dud generated art).
I need to know the environmental cost of generating an image with AI vs running a PC for hours making one by hand
Per-image impact depends on
- what kind of art device you’re using, desktop vs tablet could be a 10x difference
- how many tries it takes to get a generated image you like
I would guess under $2 of electricity either way. Overall impact will be dominated by fixed setup costs of manufacturing devices and training models
Hard to compare, but roughly 0.005-0.05 kWh for an already-trained model to run once. A small number of additional outputs simultaneously can be produced for relatively little additional cost by most models. Edits and rerunning will vary, usually lower, but can rarely be more energy intensive.
For a digital artist to produce a rendition, it will vary a ton but for a relatively efficient PC with drawing software running, maybe 70 W, that’s 0.07kWh/h (or 70Wh) however you want to think of it. So multiply 0.07*(number of hours).
I think there’s stronger arguments against image generation, for commercial enterprises at least. I don’t really draw and haven’t inspected the files from image generators to know, but without recordings of the creation process I’d always wonder if someone I don’t particularly know that I’ve commissioned is really drawing it or not. This can only constrict us further, under capitalism. Fewer people will trust each other, just use machine models for low cost. Wealth probably ends up concentrating in fewer hands from there, even if more people are generating things, they’re vying for fewer patrons capable of compensating them. There’s some carrying capacity for entertainment, which is mostly what this stuff can produce at this point. Until it’s plugged into more robots, assuming things don’t break before that.
