Raw LLM Responses

I want to say that some of the concerns here (demotivation, companies trying to use image synthesizers to replace artists, trolls feeding someone's work into a synthesizer and modifying it slightly and then trolling the artist with it) are all 100% legit and are concerns with how these tools are being used by people. The ethical questions surrounding the data-sets used to train the program are, however, a very different issue. An argument needs to be put forward that what these programs are doing isn't covered under fair use for the purpose of making transformative works. I've seen lots of insistence that these programs are in some capacity violating copyright, but I haven't seen an actual argument substantiating that claim yet. I don't think such an argument is impossible. It might be doable. But also consider for a bit what the legal ramifications such an argument would have on the creative world. If this technology isn't falling under the transformative works clause of fair use I have a really hard time imagining ANYTHING being covered. Image synthesizers are transformative on a scope and scale that's frankly difficult to wrap one's head around. We know these programs were trained in part on Disney IP and if Disney thought they had any case at all regarding copyright infringement Dall-E and Stable Diffusion and all the rest would have been nuked from orbit already. I don't think we need to give giant corporations MORE legal tools for crushing any form of competition based on spurious claims of 'malicious inspiration.'