[Coda]

I suspect that isn't true, although it may only be untrue based on a technicality -- Adobe's terms of service for their cloud offerings almost certainly include blanket permission to store, process, and analyze anything you have in your account. This is obviously necessary for them to be able to store your data for you and show you thumbnails, but it would also grant them permission to use your files in training their AIs.

But on the other hand, I'm not even sure that's unethical. That same agreement limits the manner in which they're allowed to share that data. I don't expect that Adobe's lawyers would be happy with the idea that the company might be redistributing data that their users believed was being securely and privately stored. If their AI models aren't capable of reproducing their users' original creations, then there's not really a problem, especially not a problem that one could reasonably claim damages about.



I'm also not sure that the unethical users are in fact the loudest -- I think they're just the ones being amplified the most. That is, they'd probably not seem like such a big deal if they weren't being called out. This is, of course, working as intended: the whole point is to draw attention to the problem.



It's hard to say that this is a "capitalist" thing, though. Stable Diffusion is an open-source model, so it's not making money in and of itself. The problem with stopping the unethical practices isn't a matter of there being too much money involved -- on the contrary, the big attention and the big money are more interested in PREVENTING the unethical uses. It's a matter of there not being any way to actually enforce it. Most unethical users are going to fly under the radar, just like how most software pirates (and for that matter, art thieves that don't use AI) get away with their crimes because it's impossible to track and impractical to punish.