When you sample in music, you get the original artist’s permission or you get fucking sued. If the AI used were trained on a licensed library catalogue, then sure. Media companies historically would buy sample licenses to use for their sound effects in movies, video games, etc. so AI could essentially just do that, but put the encyclopedia of samples in a blender of training to modulate that shit to make something somewhat “new” to be used. Original artists get royalties, users get something customized without having to hire sounds engineers to make those adjustments, and consumers get good products.
Yeah, that’s basically what I was trying to imagine. It’s absolutely not what contemporary AI is, but it’s closer to how I think the technology should be used.
When you sample in music, you get the original artist’s permission or you get fucking sued. If the AI used were trained on a licensed library catalogue, then sure. Media companies historically would buy sample licenses to use for their sound effects in movies, video games, etc. so AI could essentially just do that, but put the encyclopedia of samples in a blender of training to modulate that shit to make something somewhat “new” to be used. Original artists get royalties, users get something customized without having to hire sounds engineers to make those adjustments, and consumers get good products.
Yeah, that’s basically what I was trying to imagine. It’s absolutely not what contemporary AI is, but it’s closer to how I think the technology should be used.