It’s you that’s spreading misinformation. Is it possible to train an ai not to infringe copyright? Maybe. But it clearly happens today, with available models.
To add to that, I consider this to be unintentional plagiarism. As someone who is about to finish a Masters degree it is no mystery that two humans can come up with the same thing. This can constitute copyright infringement, but it is far from theft.
Many are spreading misinformation that Generative AI art is akin to copyright infringement. While this is still being disputed legally, the technical answer is no. AI works much like the creative part of our brain, getting ideas from things it has witnessed, and creating works of it's own based on that.
Many are spreading misinformation that Generative AI art is akin to copyright infringement. While this is still being disputed legally, the technical answer is no. AI works much like the creative part of our brain, getting ideas from things it has witnessed, and creating works of it's own based on that.
AI works much like the creative part of our brain, getting ideas from things it has witnessed, and creating works of it’s own based on that.
Only if it 'sees' the content in real time when programming itself, like a human being does.
But if a corporation copies that content/data to their servers/hard drives first, then runs training programs on said content/data afterwards, then its copying the content/data first, and violates copyright.
I see your point. But it's not based on anything. You just claim AI creativity works like human cretivity without backing that up. While I too think this is the case... The interesting question would be: Is that really the case? And why? Without an explanation and reasoning, it's just an unsubstanciated claim.
And it's not the whole story of what's happening in practice. Several companies have been proven to download everything they can get a hold of and train their AIs on it. Pirated book archives, images they didn't obtain rights to. To the point where watermarks crept in and they had to add steps to their procedure to remove them and hide the fact. And Meta had to admit to pirating books. A proper analogy would be a professor who uses bittorrent and pirating sites to obtain material and hand that to their students to learn from... And that's not how it works in real life. Schools and universities license the material to distribute it to their students, put it in a library or the students have to buy the books. They don't rip off school-book authors per default.
And am I allowed to run bittorrent and download the latest Marvel movie (without paying) to watch it and "learn" what it's about? Download lots of music from P2P networks like I used to in the early 2000s? It'd save me the money I currently spend on a Spotify and Netflix subscription.
I think Artist should vote for system where they don't need to fight for some otherwise unnecessary 'property rights'. Try ditching the Capitalist religion.