AI PBs
-
I have yet to hear actors actively speak out about pictures from their movies being used to represent a RP character in a tiny game.
I often wonder how much of that has to do with news coverage. Is any reliable news source regularly covering our hobby?
-
@MisterBoring said in AI PBs:
I often wonder how much of that has to do with news coverage. Is any reliable news source regularly covering our hobby?
LOL what? …no…
Anyway, if this was going to go anywhere in terms of litigation it would’ve come up in the 2000s when fan fiction and websites that did ‘dream casting’ for movies were just becoming things.
-
Image generation isn’t the only thing generative tech like this is being used for. It’s being used for animation, for voices, etc. Actors are fighting for their faces and voices to not be used as an AI simulacrum of themselves to put them out of work on a broad commercial scale. The creation of deepfakes and simulated representations of real people is another reason why I personally find generative learning algorithms deeply ethically concerning. That’s not necessarily a problem with Midjourney itself, and I can see why people might draw the line between a deepfake and a generated artistic rendering of an elf, or whatever. But there’s a reason I categorically dislike it.
-
It doesn’t come up because the impact is minimal. There’s no likelihood of confusion that it’s an authorized licensure and no one is making money on it. People who make money using their likeness know - assuming they know or care that online RP exists - that it does not cost them anything for their face to be used in this fashion because it is not an area of business that it would earn them any money. Essentially, is it morally correct to “steal” someone’s face for this purpose? Probably not, but also who cares? What harm exists here?
Basically if you want to use AI PBs go ahead, but this idea that it is somehow morally superior because no real people are involved is disingenuous. The only choices that are free of contention are drawing your own shit, paying to commission an artist to draw your own shit, or not using an image at all and just writing descriptions in this writing hobby. Otherwise we are all making choices that have points of compromise.
-
Speaking of litigation, who do I speak to about suing for the whiplash I got as those goalposts fuckin’ rocketed past?
-
You are paying for everything being above board. If the corporation is pulling funny business, then that’s on them (and Disney’s lawyers will undoubtedly made them pay).
That’s just not accurate. Tools like Midjourney provide absolutely no claims that they have the rights to the materials used in their training data or that the images they generate are safe from copyright or trademark infringement claims. Mainly because they don’t. That’s why there are eleventy billion lawsuits against GenAI companies from artists, authors, publishers, and more.
And EVEN IF the law concludes that training these GenAI systems on copyrighted material is fair use (which is still very much up in the air), that still doesn’t protect you if something you generate with them infringes on someone else’s copyrights or trademarks.
I can ask GenAI to make me an image of a golden space robot and slap it on a T-Shirt to sell, but that’s not going to save me from a DMCA takedown or lawsuit once Disney sees that it’s the spitting image of C-3PO.
@Third-Eye said in AI PBs:
Anyway, if this was going to go anywhere in terms of litigation it would’ve come up in the 2000s when fan fiction and websites that did ‘dream casting’ for movies were just becoming things.
That’s what I was alluding to when I mentioned fancasts/faceclaims not being unique to MUSHes. Like fanfic/fanart, there are certainly those who take issue with them. And like fanfic/fanart, they exist on shaky copyright ground, relying mainly on the good graces of the creators.
Even so, there never has been anywhere near the widespread outcry from creatives over these things than there has with GenAI. If you have a personal problem with them, that’s totally fine! But thinking that GenAI is better than fancasting because it somehow harms creatives less feels like it’s almost willfully ignoring what creatives themselves are saying.
-
-
If you need more proof that generative AI is super bad, take some advice from one of it’s initial investors:
https://futurism.com/openai-investor-chatgpt-mental-health
IT WILL DRIVE YOU INSANE.
-
@MisterBoring Technological Darwinism.
-
So AI isn’t so much coming for my job as it is creating more work for my colleagues. I see.
He/Him. Opinions and views are solely my own unless specifically stated otherwise.
BE AN ADULT -
@MisterBoring said in AI PBs:
Are there any other options that might represent a truly ethical source of PB art?
My tabletop players have made HeroForge minis of their PCs. Not actually downloaded them or had them 3d printed or anything, just screenshots of the miniature online. So, it’s free, and I don’t think anybody minds. Mind you they do sometimes buy stuff from HeroForge, too.
-
@Gashlycrumb said in AI PBs:
@MisterBoring said in AI PBs:
Are there any other options that might represent a truly ethical source of PB art?
My tabletop players have made HeroForge minis of their PCs. Not actually downloaded them or had them 3d printed or anything, just screenshots of the miniature online. So, it’s free, and I don’t think anybody minds. Mind you they do sometimes buy stuff from HeroForge, too.
I have also seen people use things like BG3 and Cyberpunk’s character creators in a similar fashion for VTT.
(And it’s becoming more common for games to release character creators for free in advance of the game’s release.) -
I’ve actually been looking at using Unreal Engine’s Metahuman Creator to make PBs, but my abilities in that engine are… awful.
-
So AI isn’t so much coming for my job as it is creating more work for my colleagues. I see.
here’s another article about it more generally. I find it kind of funny but also deeply sad considering that these are probably just…what, undiagnosed narcissists who have badly needed intervention for years? and now they’ve got Fancy Autocorrect endlessly validating them into psychosis.
-
@Wizz There’s a dark part of my psyche that hopes it leads one of them to this:
-
@Wizz Yeah, I was being a bit flippant, but some colleagues of mine have already reported (anecdotally, their study is as-yet unpublished) an increase in cases of delusion being “fed” by hallucinating LLMs supposedly confirming cases of unreality.
Alas, most research I’ve seen thus far has been on clinical applications of LLMs rather than their clinical impact.
-
-
@Faraday Artists can’t win in court against most a.i. cases because training an a.i. isn’t the same as copying someone’s work.
A lot of people picture a.i. art like it’s theft because they think the program is just doing something like cutting a woman out of a painting and pasting it over a picture of a mountain cabin next to a lake taken by a photographer then grabbing a picture of someone’s dog with a stick in its mouth from Facebook and slapping a filter over it to make it all the same style. That isn’t how it works though. Everything the a.i. makes is entirely original.
The training process teaches it to draw in the same way humans learn to do art, trial and error with self reinforced learning when it gets it right and tweaking it’s method when it gets it wrong. It tests to see how similar what it makes is to the original training material. The reason it’s so powerful is that it can learn 24 hours a day without humans being involved at all past the initial set up.
Nothing is actually copied. Artistic style is emulated, which can’t break copying laws.
-
@RedRocket said in AI PBs:
Everything the a.i. makes is entirely original.
GenAI makes nothing original. Every single thing it does is algorithmically based on the work it’s been trained on. Without that trained work, they’ve got no product.
That trained work was used without the permission of the creators. That is the crux of the lawsuits, and while the results have been mixed so far, I believe ultimately the creators will prevail in some form or another (probably a watered-down global licensing pool, but it’s at least something). I believe this because one of the cornerstones of the fair use doctrine is that the transformative work does not replace or compete with the original. That is demonstrably not the case here. This has been theft and plagiarism on a scale that would make Napster blush.
ETA: The Getty and Disney lawsuits are probably the strongest, as they show pretty compelling evidence that their artwork/photos are baked into these GenAI tools to such a degree that it can faithfully reproduce them when prompted. It’s not just stylistic inspiration.
@RedRocket said in AI PBs:
The training process teaches it to draw in the same way humans learn to do art…
GenAI does not learn in the same way a human does. It’s a false equivalence. People keep wanting to anthropomorphize these things like they’re actually intelligent, but they’re not. They’re fancy word- and image-predicting algorithms. Autocomplete on steroids. They do not fundamentally understand the world the way a human does. They have no actual creativity, insight, or originality. They match patterns and generate similar ones. They do it really well, which is why the tools work, but that is not the way humans think or learn.
-
Every single thing it does is algorithmically based on the work it’s been trained on. Without that trained work, they’ve got no product.
Yes, but the same is true for humans.
If you never learned to draw by trial and error, by comparing your work to other people, by learning anatomy and seeing how close you can get it to a goal you set for yourself, you wouldn’t be able to make anything either. Your brain and the A.I. brain work the same way. That’s why, legally speaking, it isn’t copying. It’s a very skilled imitation, yes, but it isn’t copying.
