Live data from Hacker News

Learnings from paying artists royalties for AI-generated art

kapwing.com

151–160 of 176 posts

Re: Learnings from paying artists royalties for AI-generated art

#151
post #3

Props for a postmortem, much like scientific studies that publish negative results.

Fairly surprised they admit the project burned out an engineer enough that they quit.

“One engineer who left Kapwing in fall of 2025 said that the short-lived Tess investment contributed to burnout.”

Re: Learnings from paying artists royalties for AI-generated art

#152

Earlier quoted context omitted.

Courts pretty much always rule in favor of rich corps that steal from individuals, and increasingly so. AI companies have money. Artists don't. That makes AI thievery fine, doubly so since AI corps have financially contributed to the government.

Look, you've made a closed argument. Now if I mention small labs or floss projects that got litigated against, first I'd need to 'stop beating my wife'. No one is stealing anything. It's not theft. There has been no crime. None of this is anywhere near criminal law. I could make a more nuanced argument on copyright infringement. But to make that steelman, I'd need to accept a too large overton window shift, so I'll d…

It can output books verbatim. It often "mistakenly" embeds watermarks from famous artists into generated pictures. Arguing that it's not stealing because a bought and owned legal system, which worked at a glacial pace even before it was completely bought off, isn't theft just because a law doesn't exist yet is silly. It's analogous to saying that dumping uranium and blowing up nukes everywhere in the 1940s and 1950s was great and non-polluting because there wasn't a law against it and nobody is being hurt (and cancer takes a while to develop, so nobody can prove their cancer was from nukes nearby). People argued back and forth back in the day about it. Now we realize that waiting for laws was dumb because it was pretty obviously bad not just in retrospect, but at the time. AI makes shoddy copies of good stuff. It pollutes the internet in ways that'll outlive us, just like a nuke does to the world. And it's pure cancer.

Re: Learnings from paying artists royalties for AI-generated art

#153

Earlier quoted context omitted.

Look, you've made a closed argument. Now if I mention small labs or floss projects that got litigated against, first I'd need to 'stop beating my wife'. No one is stealing anything. It's not theft. There has been no crime. None of this is anywhere near criminal law. I could make a more nuanced argument on copyright infringement. But to make that steelman, I'd need to accept a too large overton window shift, so I'll d…

It can output books verbatim. It often "mistakenly" embeds watermarks from famous artists into generated pictures. Arguing that it's not stealing because a bought and owned legal system, which worked at a glacial pace even before it was completely bought off, isn't theft just because a law doesn't exist yet is silly. It's analogous to saying that dumping uranium and blowing up nukes everywhere in the 1940s and 1950s…

Nope. Besides not stealing, it's also not nuclear proliferation, cancer, or pollution either. Nor are courts ever likely to call it that. Not even if the defendant is a poor European student. Especially not if the court is actually clean.

The problem is that you're putting it in the wrong legal framing, and it just won't fly. Willing to engage, but not on these terms.

Re: Learnings from paying artists royalties for AI-generated art

#154
post #150

Earlier quoted context omitted.

There are definitely >1m artists worldwide, some popular some less so, and $1M * 1m =1T, not 0.1% * 200T =200B. Hard cap of 200B divided by 1M equals 200k, and that would be sure more reasonable, but we aren't hearing artists responding favorably to hypotheticals in that range, so I'm skeptical that "ain't nobody gonna turn that down".

This isn’t ownership it’s a license. I think the vast majority would agree to let AI companies train on their art for 10k let alone 200k. Don’t forget the average global salary is way below what you see in the US. Put another way how many people would turn down 6+ months salary. Of course the vanishing tiny percentage people care about would want more, but that’s a separate question and not particularly valuable to A…

> Put another way how many people would turn down 6+ months salary.

Didn't that exact social experiment took place in the US last year? I thought the result of that was disastrous if media reports are to be believed.

OTOH I remember creator of Wordle closed the "low few mil" deal instantly, so I do believe it unlikely that people turn down few _hundred_ months worth of salary. But those artists are not from regions with 50-100x less median income and/or wider income distribution relative to US - I think they're concentrated in relatively high-income-low-disparity regions - so I don't think there's backwater wherever that lifetime income there is equivalent to no more than 6 months worth in US that has abundant supply of artists.

And IMO those artists are basically engaged in a geo-scale dumping of media contents. It's the same phenomenon as how moving consumer electronics manufacturing to US instantly multiply costs by small integers instead of just incurring premiums in percentages. If that phenomenon were to be quenched and those effects were integrated into economy anyhow, that will change the global balances of power to some statistically significant degrees, like, we'd be seeing flying rocket amphibian McBoatfaces everywhere. That might be interesting, but I'm not sure if that's an interesting kind of an interesting thing to see.

Re: Learnings from paying artists royalties for AI-generated art

#155
post #150

Earlier quoted context omitted.

This isn’t ownership it’s a license. I think the vast majority would agree to let AI companies train on their art for 10k let alone 200k. Don’t forget the average global salary is way below what you see in the US. Put another way how many people would turn down 6+ months salary. Of course the vanishing tiny percentage people care about would want more, but that’s a separate question and not particularly valuable to A…

> Put another way how many people would turn down 6+ months salary. Didn't that exact social experiment took place in the US last year? I thought the result of that was disastrous if media reports are to be believed. OTOH I remember creator of Wordle closed the "low few mil" deal instantly, so I do believe it unlikely that people turn down few _hundred_ months worth of salary. But those artists are not from regions w…

Wordle involved actually selling the rights to something not just allowing AI to be trained on it while he kept the website.

That’s really not a reasonable comparison to what is being sought.

As to global artists, I was suggesting the majority of artists globally make ~20k USD or less per year as artists. To get to millions of artists you need to use a generous definition, so now Hollywood is full of actors how many of them made 20+k last year as an actor? If you disagree fine let’s double it and 6 months salary is still only 20k and would I suspect be a seriously tempting offer when you retain all rights to past and future works.

Re: Learnings from paying artists royalties for AI-generated art

#156

Earlier quoted context omitted.

The old rules were built on based on old capabilities and and old reality which no longer exists.

This is a deceptive line of dismissal. Sound principles needs to be figured out before imposing any kind of restriction on art - "things have changed" doesn't cut it.

Figuring things out is exactly what needs to happen. I think it is valid to dismiss arguments of “this is how copyright has always worked” when those rules were written before AI completely changed the game.

Re: Learnings from paying artists royalties for AI-generated art

#157

Earlier quoted context omitted.

They took a base model, trained on but not reproducing work, so entirely fair with no theft, and then tried to tweak it so it could make money for an artist.

Except that as soon as it is used to create work, it’s reproducing work that is derived from what it was trained on. Not just the stuff it was TUNED on or asked to derive style from.

Its derivative but not to any infringing extent.

Re: Learnings from paying artists royalties for AI-generated art

#158

Earlier quoted context omitted.

>IP laws can stay the same, but they should have purchased a license to use my art before including it in their training data. But including your art in the training data is fair use (or otherwise exempt) by most standards, as no reproduction occurs. You are advocating for a change to IP law to make it more restrictive.

Fair use by most standards? Which standards are those? I don't think a standard about training an AI on billions of images exists.

Google scrapes the entire internet to generate a searchable index of the internet. But the resulting search engine is only infringing where it reproduces entire copies of scraped news articles and images. Both places where they have been put back in their place through legal means.

Like LLM's, it retains the produced index but not the original data.

The big concern is whether producing an LLM is competing with artists directly, but as artists dont make LLMs, this seems to be consistently ruled as non competing.

Re: Learnings from paying artists royalties for AI-generated art

#159
post #103
post #60

Earlier quoted context omitted.

By the same 'transformative' standards that allow satire, reaction and commentary videos to exist. And those take 100% from the source and add context, whereas good generated AI images that aren't wholesale copying take like less than 10% from the original source. In addition, the idea that you need to pay rent on *your observation* of someone else's work is absurd. No one pays Newton's descendants for making lifts o…

Is it transformative if I take all the pages in Hanya Yanagiharas A Little Life and use a thesaurus to change every second word? Or a more realistic scenario: what if I translate it to Spanish without license from the author? That's not allowed, and yet I have "transformed" the work in the same way that an LLM does.

If I buy a book entitled "How to make a table" and then make a table, the author does not own the table I made.

If I buy a book and use it to prop up a table, the author likewise does not own the table, or any works I undertake on that table.

If I buy a book and rip out the pages to make a collage, the US is the only legal jurisdiction where I run even slight risk of civil penalties.

An LLM is downstream of a book. Using a book to make an LLM does not confer any rights or privilges towards the LLM on the original author, just as using a hammer or nails dont permit the hammer or nail manufacturers any royalties on what I make, even if I build a hammer making machine with them. Theres no right to the works of people who build on your work without reproducing your work, at least outside of strict copyleft.

Its like demanding a cut from people who learned how to use photoshop by watching your photoshop tutorial youtube videos.

This is why the most successful cases against LLMs have been on the "Did they purchase the book" side of the fence, and not on the "What did they do with it" outside of the one case, where the legal company tried to use the LLM to 1:1 reproduce the content they had a limited license to, but thats obviously a no go and they should have known better.

Re: Learnings from paying artists royalties for AI-generated art

#160

Earlier quoted context omitted.

>IP laws can stay the same, but they should have purchased a license to use my art before including it in their training data. But including your art in the training data is fair use (or otherwise exempt) by most standards, as no reproduction occurs. You are advocating for a change to IP law to make it more restrictive.

> But including your art in the training data is fair use It shouldn't be!

Why? I dont undertand this take at all.
Post reply on HN