AI PBs
-
@MisterBoring said in AI PBs:
Pretending that commercial generative AI models aren’t walking databases of art theft is just disingenuous at this point.
I agree with this, but do you think that non-commercial generative AI isn’t also art theft?
I think that it is theoretically possible for it not to be. If I trained my own model on photos I took, or art I did, then that could probably be reasonable (from an art use standpoint, at least). Whereas all commercial generative models, at least so far as I am aware, are prolific in their art theft.
This is largely not directly connected to whether a model is commercial or not. A commercial model could theoretically build a dataset on only content they have a legal basis to use. A nonprofit model could build a dataset on stolen content.
I’m more comfortable using an AI generated image than the image of a real person who didn’t in any way agree to be used in that way. Same with AI generated ‘art’ (gosh what a stupid term) rather than something just sniped off the internet, give I’ve never, and probably never would, pay someone to create a personal piece just for a PB for random RP.
It’s always been some level of moral iffy to me, and the introduction of AI is just a different kind of moral iffy.
This is what I don’t get here: all sorts of artists (across all artforms, not just speaking visually) have spoken out against generative AI of different stripes. Like, this is an active battle that creatives are fighting. We know tons of creatives who publicly state they do not like generative AI.
When has there been similar commentary regarding the kind of tiny-scale hobby usage that’s been done in RP communities for decades? This is a sincere question, because it may very well exist! But to me, what I see is people saying, “I’m bothered by X thing on behalf of creatives (who have not commented on X at all) and think that Y thing (that creatives have actively and repeatedly spoken against) is better.”
When companies use AI to cheap out on hiring artists, there’s a tangible loss in the equation. The artists’ output was stolen to create a system that then squeezes the artists out of their livelyhood. I’m fully on board with that being shitty on so many levels. The same with ‘AI prompt Artists’ who are taking actual money out of the pockets of other people.
But that’s what using and popularizing the product supports. There is a direct line from A to Z here.
I’m not sure I’m more comfortable, but you touch on why I really waffle on the whole “we’re still using someone else’s copyrighted work.”
Something like 15 years ago, I took two copyrighted images of Ben Affleck and Ray Stevenson and clipped them together (very badly) so it looks like they’re kissing. I don’t think either of them would have consented to the existence of this image, and now one of them is dead, so he definitely can’t.
Today, I’d ask Midjourney to create the same image and it’d probably take about the same amount of time and probably create close to the same image I did.
I know that people are going to insist that using MJ is more exploitative of artists because it was trained on artists’ work without consent, etc., but Ben Affleck and Ray Stevenson are also artists, and I never paid them for their likenesses; the photographer who took the pictures I snipped and clipped is an artist, and I never paid them for their work; nor the websites I right-clicked to take the hosted art from to begin with…
The impact to the actor’s here seems the same or worse in the MJ example. In both examples, you’ve created an image that didn’t exist before in that form, and using the actors’ likenesses in a way they never actively consented to. But in the MJ model, you’re also engaging in popularizing systems that these same sorts of creatives will speak out against.
I don’t think MJ is more exploitative of artists. I know that artists actively say it’s exploitative of them. That it’s affecting their livelihood. This argument wouldn’t drive me so nuts if it wasn’t full of people saying that something creatives are actively speaking out against is less harmful than this other thing that none of them seem to mention.
-
Yeah like… this isn’t so much about the IP implications of this use case of the tool – I do not think anyone on God’s green earth really cares where we get the pictures we use to supplement our games of let’s pretend as long as we make no money. It is the fact that the tool itself was trained in a violative way that makes it gross. Taking pictures of actors and going “now kiss” is something you can do without impact, but using a tool to do it that can only do it because it was fed a bunch of stolen work is qualitatively different.
BUT I don’t think anyone HAS to care about it. I get that there is a limited supply of fucks to give and God knows i seem to have a lessening supply daily.
-
If you want to boycott Amazon…
I do this. It was a huge inconvenience at first, as it’s really hard to find anything for sale online not through Amazon. Like, I’ve even bought things off Etsy that turned out to just be repackaged Amazon stuff.
I don’t judge MUSHers who still use AWS or bought their keyboard off Amazon just because I choose not to engage with that company. Not everyone can or should make the same choices I do.
We can each choose what causes are important to us.
And how we engage with people who align differently.
On Dragon Wings · https://pern.gaslightswitch.com · pern.gaslightswitch.com port 4201
-
I do this. It was a huge inconvenience at first, as it’s really hard to find anything for sale online not through Amazon. Like, I’ve even bought things off Etsy that turned out to just be repackaged Amazon stuff.
Not to hugely sidetrack this thread, but I am also an Amazon boycotter. Most of the things you find on Amazon AND on Etsy can usually be found on Ali Express and other similar places because they all come from various dropshippers overseas. When that’s the case, I usually go straight to AE. I find it really exasperating how hard it can be to find legit products on Etsy sometimes.
she/they
-
Honestly I find that so valid. I haven’t managed to de-enmesh myself from amazon even though I 100% believe it would be the morally correct choice.
-
@Tez ooooh! Thank you for the tip! Added to the knowledge bank.
It’s a jungle out here. (get it??? cuz the Amazon is like a jungle and the company is called…)
-
And how we engage with people who align differently.
Sure. But there’s a big difference between “I understand the harm that Amazon does but I still choose to use it because (reasons, which may even be wholly justified given your personal situation)” and “I don’t get why everyone keeps saying Amazon is a big deal; it really does no harm when I order from them; people are just overreacting.” I see a LOT of the latter when it comes to GenAI, and that is what I push back on. (not from you specifically, just in general)
Like, piracy sites actively harm authors on a large scale. You can argue “I wouldn’t have bought the book anyway so I didn’t personally do any harm”, but that’s discounting the real harm caused by the very existence of those sites. (including that the pirated material was then used to train GenAI, bringing us full circle…)
-
And how we engage with people who align differently.
Sure. But there’s a big difference between “I understand the harm that Amazon does but I still choose to use it because (reasons, which may even be wholly justified given your personal situation)” and “I don’t get why everyone keeps saying Amazon is a big deal; it really does no harm when I order from them; people are just overreacting.” I see a LOT of the latter when it comes to GenAI, and that is what I push back on. (not from you specifically, just in general)
Like, piracy sites actively harm authors on a large scale. You can argue “I wouldn’t have bought the book anyway so I didn’t personally do any harm”, but that’s discounting the real harm caused by the very existence of those sites. (including that the pirated material was then used to train GenAI, bringing us full circle…)
Yeah. I’ve only been motivated to post because in response to specific points or arguments that I find unsubstantiated. (Or, in the instance of one poster claiming that MJ’s training data was all licensed, flat out wildly false.)
-
Agree w/ Faraday and Roz. We live in a world that forces us to make choices and compromise every day of our lives to try to do the most good, or at least the least harm, all while knowing that there are vast organizations constantly enacting destruction on scales we could never reach. We are only human and we have to live. We will never be perfect.
But at least make those choices with open eyes.
-
But there’s a big difference between “I understand the harm that Amazon does but I still choose to use it because (reasons, which may even be wholly justified given your personal situation)” and “I don’t get why everyone keeps saying Amazon is a big deal; it really does no harm when I order from them; people are just overreacting.”
There’s also a big difference between “I don’t purchase things from Amazon” and “I feel that people who purchase things from Amazon are enabling corruption and exploitation.”
And I think that’s why you have people responding the way they are. The upstream problem is Midjourney, but there’s a sense in this thread that players who use it are somehow to blame for that corruption.
people with AI PBs : steal from artists :: people who shop on Amazon : run a sweatshop
Midjourney : steals from artists :: Amazon : runs a sweatshop
It may not be intentional, but that’s how it comes across, and why I personally felt the need to defend my use of image generators.
I’m not a bad person! I just can’t do art and get bored!
Edit to add - @sao enmeshed is a great way to put it. They’re so everywhere that even actively trying to avoid them doesn’t always work.
On Dragon Wings · https://pern.gaslightswitch.com · pern.gaslightswitch.com port 4201
-
Honestly, I don’t think the vast majority of people care where someone’s PB comes from so long as it’s a reasonable image of the supposed character…and it’s hot enough to want to fuck, let’s be real.
I don’t think less of anyone who uses a Midjourney PB, and I’m pretty irritated that such a cool tool has been set up in a way that fucks over a lot of artists instead of licensing material or paying royalties or something. I don’t see it as a Big Ethical Question–I just wish the companies in question would be forced to pay the artists for the training data they’re profiting off of.
-
Honestly, I don’t think the vast majority of people care where someone’s PB comes from so long as it’s a reasonable image of the supposed character…and it’s hot enough to want to fuck, let’s be real.
I don’t think less of anyone who uses a Midjourney PB, and I’m pretty irritated that such a cool tool has been set up in a way that fucks over a lot of artists instead of licensing material or paying royalties or something. I don’t see it as a Big Ethical Question–I just wish the companies in question would be forced to pay the artists for the training data they’re profiting off of.
You got a +1 for the “hot enough” aside, for the record.
More to the point, namely your last sentence in particular, is where I’m at right now (of due compensation being a necessity here).
-
But there’s a big difference between “I understand the harm that Amazon does but I still choose to use it because (reasons, which may even be wholly justified given your personal situation)” and “I don’t get why everyone keeps saying Amazon is a big deal; it really does no harm when I order from them; people are just overreacting.”
There’s also a big difference between “I don’t purchase things from Amazon” and “I feel that people who purchase things from Amazon are enabling corruption and exploitation.”
I mean…presumably if a person has made the choice to not purchase things from Amazon, it’s because they identify some sort of harm in supporting Amazon’s model? If there was no perceived harm in using it, then they wouldn’t feel it necessary for themselves to personally abstain.
-
I don’t like using AI PBs because they all have the same face. Especially female images.
-
There’s also a big difference between “I don’t purchase things from Amazon” and “I feel that people who purchase things from Amazon are enabling corruption and exploitation.”
I purchase from Amazon and thereby enable corruption and exploitation. I have reasons, but they’re kinda selfish. I can at least admit it.
Someone using GenAI tools is supporting a tool that steals from artists. Full stop. You can argue that your support is a drop in the bucket (as someone can for Amazon), but it is undeniably contributing to that bucket. Every GenAI query harms the environment more than its alternatives. Every GenAI query is a number reported on a spreadsheet of “look how many users we have!” that is used to justify more corporate investment in tools that harm artists.
Where I draw the line is “people who ((do thing I disapprove of))” are evil/bad/scum of the earth/deserve harm/etc. That’s going too far, and I feel it can be avoided with a tiny bit of empathy.
-
There’s also a big difference between “I don’t purchase things from Amazon” and “I feel that people who purchase things from Amazon are enabling corruption and exploitation.”
I purchase from Amazon and thereby enable corruption and exploitation. I have reasons, but they’re kinda selfish. I can at least admit it.
Someone using GenAI tools is supporting a tool that steals from artists. Full stop. You can argue that your support is a drop in the bucket (as someone can for Amazon), but it is undeniably contributing to that bucket. Every GenAI query harms the environment more than its alternatives. Every GenAI query is a number reported on a spreadsheet of “look how many users we have!” that is used to justify more corporate investment in tools that harm artists.
Yes. There’s a difference between “I use this knowing the tool is exploitative, but I really don’t think my tiny usage is going to make a real difference” and “I use this because I think it’s less harmful.”
-
I’m thoroughly confused by what the argument is at this point. Is it ok to use AI as long as i acknowledge it’s harmful to artists?
-
@bear_necessities I’ve been arguing primarily in response to the posited idea that generative AI is less harmful or objectionable to the involved creatives than using existing imagery.
I don’t like the growing prevalence of AI imagery in the hobby, but that’s not a crusade I’m particularly willing to take on, and hasn’t been the point of my arguments. I’m just arguing about the framing that’s been centered on “AI is less exploitative to the creatives.”
-
The problem with that theory is that most of the current generation of LLMs only work at scale. Unless you had a gazillion of your own photos, or had written an entire series of novels, it’s unlikely that you could train an AI model just on your own work and have it work effectively.
I don’t think it’s possible even then. Even if you it was solely trained to replicate a single artists style, I believe we’d only end up with something that worked within the limited variety of the artist’s work, and attempting it to produce something outside of that would produce non-useful, though perhaps morbidly entertaining results. There would still exist a need to train it on a library of stock images to allow it a wider breadth of subject matter.
For example, if we fed an LLM only the entire catalog of Van Gogh’s art, and asked it to produce an image of an sports car, it would probably either fail to produce anything, or just produce something random and call it a ‘sports car’.
-
@MisterBoring That’s an interesting point. There do exist models for SD that are fed on images in public domain, but I’m curious how well they really hold up because of the sorts of examples you mention.
@bear_necessities said in AI PBs:
I’m thoroughly confused by what the argument is at this point. Is it ok to use AI as long as i acknowledge it’s harmful to artists?
It’s okay to use AI. Maybe it’s not perfect, but fuck it. I use AI, but I acknowledge the technology has some real flaws, and I don’t try to pretend that it is ethically better than alternatives. I’ve used AI for images and AI for code. I’ve even used AI to help me figure out why an update failed for BMD, so abandon ship if that’s a problem!!
I use it. I do sometimes think about whether I should buy carbon credits or something to feel like I use it ethically, but on the other hand, I don’t worry about the carbon credits I burn playing video games. I don’t know. On my fucks given scale, it doesn’t really rate, but it does sometimes itch.
In this discussion, I find the approach that AI PBs are ethically preferable to using PBs of existing persons hard to swallow.
What is the line of thinking? Many of these models have used those very same images in their training data. Like, you’re just using the exact same images with an extra layer of ‘and also other copyrighted works’, in a way that is still very much under debate for how much actual harm it causes.
Then there are some harmful beliefs out there which make people blind to potential issues:
Secondly, if the model is made for a for-profit system like Midjourney, then they already have the requisite rights and permissions. That’s part of what you’re paying for when you buy a license for Midjourney.
That’s just incorrect. Make your judgments on the matter based on fact, at least.
she/they
