Raw LLM Responses

Great video, there are little more information needed. Let's start with the law and how the lawsuits against AI companies aren't going in favor of artists in USA. The problem with the term theft is that AI, technically speaking, doesn't infringe copyright. A copyright infringement is when a copyrighted work is reproduced, distributed, performed, publicly displayed, or made into a derivative work without the permission of the copyright owner. To train AI you don't actually do none of those, you show images to it and, how it's show in the video, by the diffusion process, it tries to reorganize noise into a image. The problem here is, if we gave AI a biological body with eyes, the results would be the same. The argument that AI companies use to defend AI is, even if it doesn't learn as a human, it needs information to learn (how humans do), which means that AI doesn't use images to train exactly, only the information of patterns and styles on they (it's even destroys the original material before the training process). So, if you make a image public, anyone could use it to train AI without your consent, because you can't hold copyright on information, patterns or ether your style. Only if someone uses AI to create a derivative work of your image, that would be a copyright infringement. Of course we are talking about copyright. Privacy and others types of legislation need to have a separate discussion.