I tend to frequent rp and writing forums where people often use image generators to create images for their OCs and stories. None of it is for profit, it's all hobbyist casual stuff. Sometimes these folks will post these generated images and claim it as their "art" while acknowledging openly they have used programs to create these images. Like "look at this AI art I made".

I don't like it. Ultimately, my problem with image generators is the theft which matters primarily because those artists and creators deserve to be paid for their hard work, experience, and skill. And they deserve to have their consent respected and their rights over their own work acknowledged.

But when nobody is making money off of the images and it's just being bandied around in a writing/gaming setting, I find my stance is a bit more emotionally biased.

What do you think about this? How would you articulate an argument against casual use of image generators? Or wouldn't you?

Speaking personally, and not as site admin, I'm against it as both a for-profit endeavor or casually. If I walked into your house and started pocketing things, you'd be upset with that and would want reparations regardless of if I were stealing them for me, or to sell. This is the same thing. I feel that not decrying casual use lends legitimacy to it, and fails to educate those who might be ignorant of the ethical issues behind AI. Will there be people who don't care about the ethics of it because it's self-serving? Yes, of course. But that doesn't mean it should get a pass.

Then there are the people who are trying to claim that using AI to do art is the same thing as an artist learning to draw by imitating an artist's style or copying an image (things that many novice artists do). But, that's not the same thing. An artist learning by mimicking another artist is still learning lessons and skills by doing so, and still has produced their own piece of art. Is it derivative? Yes, it is. But it's also the artist's own sweat and time that made it. Giving an AI a prompt is nothing close to the same thing.

I'm probably the most passive or least passionate either way on the issue, but a deed or misdeed is still a deed or misdeed whether or not it's casual, so that doesn't change much. The gradience of derivation is unchanged.

"Do, or do not, there is no try." ~ Yoda

One of my gripes with personally using AI to draw for absolutely anything was: using AI to draw more obscure characters can make them looks bootleggish

If you ever feel bad about the way you draw or write then just remember that you made it all on your own without resorting to using AI.

My simple argument (if it can be seen as so) is that AI "artists" are not real artists.

As I've written before, the biggest problem isn't AI-gen itself, but how their libraries are assembled: if people could be assured of an open-source database of willing contributors, it could basically serve as a crowdsourced evolution of the paperdoll Flash programmes of yore (potentially with the added benefit of being able to trace back to specific sources in the diffusion). I have a few HoI4 mods in the works, and I'll admit, AI-gen would be a godsend for making generic leader portraits that aren't just shooping real people, but without a resolution to the Copyright Issue there's no way I can approach it in good conscience, and at this point Stable Diffusion and OpenAI may have permanently poisoned the well.

Otherwise, basically seconding Bad Karma. The punchline to the AI "revolution" is today's luxury is tomorrow's expectation—there's a reason the Luddites are a hot topic again.

I'm still immensely proud of this comic. Rather apropos, the keyword was "Original Character Donut Steel". :^)