It's hard to argue that AI companies aren't making use of copyrighted images to operate a business generating vast quantities of unlicensed derivative images th

It's hard to argue that AI companies aren't making use of copyrighted images to operate a business generating vast quantities of unlicensed derivative images that are often very close to the originals. And, on the other hand, it also seems very plausible that systems like Midjourney have substantial non-infringing uses, generating images which are different enough from the original source material. But I can't escape the feeling that regardless of how this case is resolved, all of us (not to mention independent artists whose work is also part of the training data) will end up losing.