I may be in the minority here when I say I don’t see the problem. AI trained on millions of publicly available images used to speed up the concept stage of development seems like fair use to me. Like the developer says, commercial artists have always used other folks work to speed up their development, that sounds more problematic to me than drawing inspiration from a huge dataset.
“Fair use” has a specific meaning in copyright law. If something replaces the need for something else in the market, it’s almost certainly not fair use. Generative AI replaced the need to hire an original artist.
Copyright is about specific works, not vague ideas or styles. You can’t just claim that work X violated the copyright of work Y because it has some similarities and competes with it in the market place. You have to show that work X copied substantial parts of Y. And with AI models that’s going to be difficult, as the average image contributes about a single byte of information to the model. When the model was properly trained, without excessive duplicates, there is no way to get back to the original image from the model (some exceptions do exist here, e.g. Mona Lisa).
It also worth pointing out that in all these months of discussion, nobody ever managed to show a single image that the AI model would violate the copyright of. If AI stole your stuff, it shouldn’t be that hard to find some evidence for that.
Copyright can be violated even if your output does not contain a copy.
For example, if I burn a copy of your Disney DVD, watch it, and then write a review, then I’ve violated copyright. The review doesn’t violate copyright, but the DVD I burned does. Even if I throw away my DVD after publishing my review.
All the major AIs were trained with images that were downloaded from the web. When you download something from the web, you do not have an unlimited license to do what you want with your download. You may have a right to view it, but not use it for commercial purposes such as AI training. And if you use that image for AI training without permission, then you’ve violated copyright. Even if you delete the image after you’re done training your AI.
For example, if I burn a copy of your Disney DVD, watch it, and then write a review, then I’ve violated copyright. The review doesn’t violate copyright, but the DVD I burned does. Even if I throw away my DVD after publishing my review.
No, you haven’t. Making private copies is completely fine under copyright, that was decided back in the VHS days. You might violate DMCA, but that’s not an issue for AI, as everything they were trained on was publicly available and unencrypted.
When you download from the Internet, the server makes the copy, not you.
Your review still didn’t violate copyright. You are even free to include some images of the movie in your review under Fair Use, as long as they are small and insubstantial enough to not stop people from seeing the movie.
Watch any art tutorial, step one is gathering reference images from the Internet. If that would violate copyright, than a lot of artists would be in big trouble.
The SCOTUS ruled that VHS could legally be used to time-shift TV broadcasts, ie record a program in order to personally watch later. If you have permission to watch a TV program, then watching it at a different time has no economic impact and is fair use. Making a copy of someone else’s DVD is still illegal. So is giving your VHS tape to someone else. They are not fair use.
It is illegal to download copyright protected works. That applies to the person who receives the download, even if lawsuits tend to target those who share the file.
It’s true the review itself doesn’t violate copyright, but my actions prior to the review (copying someone else’s DVD) did. It’s no different than sneaking into a movie theater in order to write the review. Focusing on the review misses the point
Any copyright protected work you gather from the Internet has a limited license. That license generally allows private non-commercial use, so most people are not in trouble.
There was actually a lawsuit by Facebook against a company that was using a web scraper to gather data about Facebook users to build advertising trackers. The judge noted that if the web scraper was downloading user photographs and text posts then it was very likely infringing IP (but not Facebook’s IP, because the rights still belonged to the users).
I may be in the minority here when I say I don’t see the problem. AI trained on millions of publicly available images used to speed up the concept stage of development seems like fair use to me. Like the developer says, commercial artists have always used other folks work to speed up their development, that sounds more problematic to me than drawing inspiration from a huge dataset.
“Fair use” has a specific meaning in copyright law. If something replaces the need for something else in the market, it’s almost certainly not fair use. Generative AI replaced the need to hire an original artist.
Copyright is about specific works, not vague ideas or styles. You can’t just claim that work X violated the copyright of work Y because it has some similarities and competes with it in the market place. You have to show that work X copied substantial parts of Y. And with AI models that’s going to be difficult, as the average image contributes about a single byte of information to the model. When the model was properly trained, without excessive duplicates, there is no way to get back to the original image from the model (some exceptions do exist here, e.g. Mona Lisa).
It also worth pointing out that in all these months of discussion, nobody ever managed to show a single image that the AI model would violate the copyright of. If AI stole your stuff, it shouldn’t be that hard to find some evidence for that.
Copyright can be violated even if your output does not contain a copy.
For example, if I burn a copy of your Disney DVD, watch it, and then write a review, then I’ve violated copyright. The review doesn’t violate copyright, but the DVD I burned does. Even if I throw away my DVD after publishing my review.
All the major AIs were trained with images that were downloaded from the web. When you download something from the web, you do not have an unlimited license to do what you want with your download. You may have a right to view it, but not use it for commercial purposes such as AI training. And if you use that image for AI training without permission, then you’ve violated copyright. Even if you delete the image after you’re done training your AI.
No, you haven’t. Making private copies is completely fine under copyright, that was decided back in the VHS days. You might violate DMCA, but that’s not an issue for AI, as everything they were trained on was publicly available and unencrypted.
When you download from the Internet, the server makes the copy, not you.
Your review still didn’t violate copyright. You are even free to include some images of the movie in your review under Fair Use, as long as they are small and insubstantial enough to not stop people from seeing the movie.
Watch any art tutorial, step one is gathering reference images from the Internet. If that would violate copyright, than a lot of artists would be in big trouble.
The SCOTUS ruled that VHS could legally be used to time-shift TV broadcasts, ie record a program in order to personally watch later. If you have permission to watch a TV program, then watching it at a different time has no economic impact and is fair use. Making a copy of someone else’s DVD is still illegal. So is giving your VHS tape to someone else. They are not fair use.
It is illegal to download copyright protected works. That applies to the person who receives the download, even if lawsuits tend to target those who share the file.
It’s true the review itself doesn’t violate copyright, but my actions prior to the review (copying someone else’s DVD) did. It’s no different than sneaking into a movie theater in order to write the review. Focusing on the review misses the point
Any copyright protected work you gather from the Internet has a limited license. That license generally allows private non-commercial use, so most people are not in trouble.
There was actually a lawsuit by Facebook against a company that was using a web scraper to gather data about Facebook users to build advertising trackers. The judge noted that if the web scraper was downloading user photographs and text posts then it was very likely infringing IP (but not Facebook’s IP, because the rights still belonged to the users).
That would be relevant if we were talking about copyright law at all.
Parent comment was literally referring to fair use.
I don’t think they were referring to fair use in copyright law. Just that it’s fair to use.