The general consensus here is that if you generate AI art at all, regardless of whether you use it commercially or not, you are engaging in art theft and are in fact an asshole.
So why doesn’t that logic get applied to straight up turning someone’s digital art and/or photos into memes and having millions of people repost it with zero attribution? I’m not talking about things like wojaks or rage comic characters where the creator intended for it to be a meme and knew for a fact that other people will copy it, nor am I talking about screenshots of popular media franchises, but the random art and photos people post that just happens to resonate with the internet in a way the creator never foresaw, becoming memes without the creator even initially knowing. Think the original advice animal meme templates like Scumbag Steve or Bad Luck Brian where it’s literally just a random photo of someone, probably taken off their personal social media. Or the two serious and one goofy dragon drawing and others that were very clearly posts on art sharing sites that got reposted with new context. I’ve even seen some meme templates go out of their way to crop out names and signatures that the original creator put there so they are credited when their work is reposted. And no one slamming AI art seemingly has a problem with any of it. In fact, if you as the creator of an image tried to get the internet to stop using your personal work as a meme with no attribution, you’d be ignored at best and targeted for doxxing and harassment at worst for spoiling their fun, probably by some of the same people condemning the use of AI.
If you go on art sharing sites, the consensus among the artists themselves is that you’re not supposed to repost their work at all unless given a CC license or otherwise explicit permission. Whether it’s for commercial use or just as a random internet post doesn’t seem to change their stance in the slightest. This implicitly includes not just AI but memes as well, as in both cases you are taking someone else’s work and redistributing it without permission or attribution. So why is this okay if AI art is not? It’s even more blatant than AI because it’s not just stealing tons of people’s work, blending them all in a neural network, and spitting out a “new” work that still has fragments of the stolen work, it straight up IS just stealing a specific person’s specific work, full stop. I feel like the reason is circular, it’s okay because it’s been happening since forever and that’s what makes it okay. And AI art is not okay because it’s new and doesn’t already have a history of everyone doing it.
The majority of people condemning AI art are not themselves artists but cite things like “respect for artists” as a reason for condemning it. But most artists aren’t just against AI but against their art being reposted by anyone for any purpose, profit or otherwise. Even if they were never going to make money from that piece, they are still against reposts on principle while most of the non-artists seem to only talk about AI separating artists from revenue. So if we’re actually to respect artists, wouldn’t we adopt that stance for everything and not just commercial use or AI?
And if this is okay, what about AI art makes it different enough to not be okay?
Finally, it’s not like people never make money off memes so a binary “AI is for profit while memes aren’t” doesn’t work.
Not trying to defend AI art, but trying to go further with the discussion that has appeared around it and genuinely trying to tease out some consistency and fundamental values in subjects everyone ostensibly feel extremely strongly about and are not willing to budge.
I’ve been trying to think of a way to have a nuanced discussion about art and AI, but hesitant to do so since people online can get incredibly vitriolic about it if they suspect that I’m in favor of AI. For the record I dislike AI being shoved in my face, from the uncanny-valley line art to the shovelware game apps that I keep seeing ads for.
To get to the point, I’m struggling to understand the moral difference between an AI model scraping art to use as training data, or an aspiring artist studying it to learn how to make it themselves. In both cases, the original artist has posted their work to a public website fully knowing that people will look at it. What’s the difference between an AI model vs a human learning from it? And how is it theft when the product is demonstrably different from the original?
I want to be clear that I just want to understand the logical underpinning of the arguments, and not arguing for one side or the other.
I know a few artists and get their complaints against AI, but I feel they’ve been way too overblown.
I look at AI as what it is - a new technology. Everthing was one at some point.
For example - cameras. Do you think artists who learned painting were happy when cameras started displacing them?
Of course there was outrage. It’s natural to protect your interests. However, technology has to be allowed to progress and people’s rights have to be respected. Developments in technology such as photography or AI are a disruption of the existing legal framework, and the two sides’ rights (those of the users and if those displaced) must be balanced.
However, unlike photography, there’s a clear legal basis and precedent analogous to AI art - in most places recieving copyrighted material without permission isn’t punishable while distributing it to others is.
An AI model is in essence a retrieval system in the sense of the US DMCA. Most other places have substantially similar laws in spirit, and most places draw the distinction between distribution and “fair” uses of infinging material. A good rule of thumb is that selling access is a big no-no, distributing is a big risk, and merely using a much smaller one. All technically illegal (as are memes).
There are at least two discussions going on here simultaneously. Is the process of a beefed up spell checker sucking up all the data the same as an artist looking at what had come before, before either of them churn out new art? I’m inclined to agree with you; the process does seem similar enough. The difference remains that one is a statistical model and the other is a human being. So even if the process appears similar enough, they are two different types of player and I can also agree that we should not treat them the same. One is able to throw constant massive amounts of spaghetti at the wall as long as there are chips and power and the other is limited by their health and more limited processing power. So where the compromise lands in this discussion simply isn’t clear yet. And while you and I can discuss this, I can say for myself at least I’m not smart enough to see where this goes eventually.
The other discussion is how all of it collides with existing copyright/trademark law, which is essentially different in every country. Constitutional rights, like freedoms of expression and the arts, are given to real people, not computers. But at least one supreme court in this planet has made corporate money a form of free speech. So eff knows where LLMs end up.
This is new territory we’re in. And I fear that’s why it will take another decade until we get a legal landmark decision or a political compromise that will be similar enough all around the world.