You wouldn't steal a handbag.
You wouldn't steal a television.
You wouldn't steal a DVD.
Downloading someone's content for AI training is stealing.
Stealing is a crime.
really, really happy that someone is calling out data-harvesting for what it really is.
The difference is: You're making copies of something:
Scenario 1: I take your baguette. Your hand is empty. You starve.
Scenario 2: I take your baguette recipe. You still have a baguette recipe. You continue to live.
Scenario 3: I take your baguette recipe and publish it. Your customers leave you. You starve.
Copying someone's IP can also impact you economically if your financial model depends on you being the only distributor of copies of something.
Should we enforce the protection of people's right to have monopoly of distribution of intellectual property?
Or should we accept that in reality, copies are free and distribution monopolies only exist in inefficient markets?
It seems totally right to protect people's intellectual property.
But information wants to be free.
It's a dilemma. Do we side what feels right, or what's real?
I am still torn on this issue. On the one hand, it feels like a copyright violation when other people's works are used to train an ML model. On the other hand, it is not a copyright infringement if I paint a picture in the Studio Ghibli style myself. The question is whether removing a ‘skill requirement’ for replication is sufficient grounds to determine a violation.
I am mostly ok with these copyright crackdowns in AI in the spirit of - if a human were to do it commercially, it would be illegal.
We can argue if that should be the case or not, which is a different issue.
However, it should not be legal to automate something at scale that is illegal when done by an individual human. Allowing it just tips the scale against labor even more.
You are not a for profit software product. You are a human. If you make a drawing, that drawing MIGHT be a for profit product.
If I make a for profit AI, that AI is a product. And if that product required others' copyrighted works to derive it's deliverable, it is by definition creating derivative works. Again, creating a small human, not creating a product. Creating a for profit AI, creating a product.
If my product couldn't produce the same output without having at some point consumed the other works, I've triggered copyright concerns.
If I make and train a small human, I am not creating a for profit product so it doesn't come in to play at all. The two are not similar in any way. The human is not the product. If THEY create a product later (a drawing in this case) then THAT is where copyright comes in.
The difference is that you’re not making money off your pictures (or else you'd probably need a licensing agreement), whereas the AI companies are making money off of the IP holders.
If you paint a Studio Ghibli totoro on your cup, pillow, PC, T-shirt, nobody is going to care. If you do this a thousand times it obviously is an issue. And if you charge people to access you to do this, it is also obviously an issue.
It's fair use if you do it, it's stealing if someone else does it to you, this is the normal playbook of the likes of Meta, Microsoft, OpenAi and the others.
Human learning is transformative. It takes time. Time, not money, is the criteria for value to human life. Humans can re-draw in styles they see and the law sees it as transformative work. Humans draw conclusions from learning and apply it to other fields - that’s natural intelligence. If AI helps humans make more new, transformative work, not the same characters as the studios, not the same plots, then the studio should not try to rent-extract. Otherwise someone will try to do it to stick figures and any other style and you’d have to teach toddlers rules for any stile before you give them a crayon. The only real impact will be that hype and styles “in fashion” will just play out faster and in the mean time the Studios get their name sealed in human minds for anything done in their style - which is more immortality than abuse.
In the world of tattooing, it’s frowned upon for a tattoo artist to take another tattoo artist’s original work and replicate it without permission, yet it’s common practice to take well known IP (Pokémon, Studio Ghibli, etc) and tattoo that on a client. The ethical boundary seems to be between whether the source artwork was created by an individual vs. a corporation.
Technically the individual tattoo artist needs a license to reproduce the IP (unless the studio released the likeness into the public domain), but it's small potatoes and these IP holders know that the free promotion helps them infinitely more than whatever they'd get from trying to enforce IP licenses.
But the AI companies are making billions off of this (and a lot of other) IP, so it totally makes sense for the IP holders to care about copyright protection.
This will be litigated and I have a feeling OpenAI/Anthropic/Claude/MistralAI will win, since we've been down similar roads before[1]. With that said, AI slop will never be a replacement for human creativity, and while AI is pretty incredible technology, I'm actually way more bullish on people.
Interesting to see how people feel about copyright law. If it’s a them the individual downloading IP it’s screw those evil IP holders. If it’s OpenAI we’ll screw OpenAI!!
I for one am interested to see how courts globally figure out what is acceptable or not. The problem imo is the cat is already out of the bag. No government is going to stop progress in LLM. I also have to wonder if we are better off in society to allow this type of IP infringement, not thinking about copying art style but being able to ingest research and other important pieces of information.
It makes wonder what will happen when robots will really take care of everything unpleasant there might be to do in life. All these laws and social systems designed to pretend merit and hard effort is why there are ridiculously disparate of wealth among humans, it's not going to accept that wage slavery is actually not only irrelevant but far more inefficient. It's not going to stop without change resistance for sure.
I think AI seriously breaks down the “transformative work” condition used to determine fair use. The work is clearly being transformed, but it also doesn’t seem fair.
These models critically depend on many hours of artistic effort during training/prompting, but don’t require any additional effort by the new distributors/consumers. On top of that, proper attribution is often impossible.
"In 1710, the British Parliament passed a piece of legislation entitled An Act for the Encouragement of Learning. It became known as the Statute of Anne, and it was the world’s first copyright law. Copyright protects and regulates a piece of work - whether that's a book, a painting, a piece of music or a software programme. It emerged as a way of balancing the interests of authors, artists, publishers, and the public in the context of evolving technologies and the rise of mechanical reproduction. Writers and artists such as Alexander Pope, William Hogarth and Charles Dickens became involved in heated debates about ownership and originality that continue to this day - especially with the emergence of artificial intelligence. With:
Lionel Bently, Herchel Smith Professor of Intellectual Property Law at the University of Cambridge
Will Slauter, Professor of History at Sorbonne University, Paris
Katie McGettigan, Senior Lecturer in American Literature at Royal Holloway, University of London.
Is it possible that the primary liability for OpenAI is trade dress? If you can produce things in (for example) the style of a Studio Ghibli film, such that an ordinary consumer can’t tell if the source is Studio Ghibli or AI, is that actionable? I feel like I see copyright concerns all the time with AI but rarely is trademark discussed.
OpenAI has ridiculous guardrails for illustrations covering any public domain subject that has been covered by Disney or any other major public corporation.
So by that benchmark Japanese companies have a case.
Try generating a 19th century style illustration of Snow White. You can't at least not on OpenAI platform.
Try generating a picture "of flying boy fighting a pirate on a ship".
28 comments
[ 2.9 ms ] story [ 54.4 ms ] threadhttps://arstechnica.com/gadgets/2025/04/you-wouldnt-steal-a-...
The difference is: You're making copies of something:
Copying someone's IP can also impact you economically if your financial model depends on you being the only distributor of copies of something.Should we enforce the protection of people's right to have monopoly of distribution of intellectual property?
Or should we accept that in reality, copies are free and distribution monopolies only exist in inefficient markets?
It seems totally right to protect people's intellectual property.
But information wants to be free.
It's a dilemma. Do we side what feels right, or what's real?
We can argue if that should be the case or not, which is a different issue.
However, it should not be legal to automate something at scale that is illegal when done by an individual human. Allowing it just tips the scale against labor even more.
If I make a for profit AI, that AI is a product. And if that product required others' copyrighted works to derive it's deliverable, it is by definition creating derivative works. Again, creating a small human, not creating a product. Creating a for profit AI, creating a product.
If my product couldn't produce the same output without having at some point consumed the other works, I've triggered copyright concerns.
If I make and train a small human, I am not creating a for profit product so it doesn't come in to play at all. The two are not similar in any way. The human is not the product. If THEY create a product later (a drawing in this case) then THAT is where copyright comes in.
If you paint a Studio Ghibli totoro on your cup, pillow, PC, T-shirt, nobody is going to care. If you do this a thousand times it obviously is an issue. And if you charge people to access you to do this, it is also obviously an issue.
But the AI companies are making billions off of this (and a lot of other) IP, so it totally makes sense for the IP holders to care about copyright protection.
[1] https://en.wikipedia.org/wiki/Sony_Corp._of_America_v._Unive....
I for one am interested to see how courts globally figure out what is acceptable or not. The problem imo is the cat is already out of the bag. No government is going to stop progress in LLM. I also have to wonder if we are better off in society to allow this type of IP infringement, not thinking about copying art style but being able to ingest research and other important pieces of information.
These models critically depend on many hours of artistic effort during training/prompting, but don’t require any additional effort by the new distributors/consumers. On top of that, proper attribution is often impossible.
From its title
"In 1710, the British Parliament passed a piece of legislation entitled An Act for the Encouragement of Learning. It became known as the Statute of Anne, and it was the world’s first copyright law. Copyright protects and regulates a piece of work - whether that's a book, a painting, a piece of music or a software programme. It emerged as a way of balancing the interests of authors, artists, publishers, and the public in the context of evolving technologies and the rise of mechanical reproduction. Writers and artists such as Alexander Pope, William Hogarth and Charles Dickens became involved in heated debates about ownership and originality that continue to this day - especially with the emergence of artificial intelligence. With:
Lionel Bently, Herchel Smith Professor of Intellectual Property Law at the University of Cambridge
Will Slauter, Professor of History at Sorbonne University, Paris
Katie McGettigan, Senior Lecturer in American Literature at Royal Holloway, University of London.
https://youtu.be/X9RYuvPCQUA?si=XXJ9l7O4Y3lxfEci
[video]
So by that benchmark Japanese companies have a case.
Try generating a 19th century style illustration of Snow White. You can't at least not on OpenAI platform.
Try generating a picture "of flying boy fighting a pirate on a ship".