Blue
Ravenous Bugblatter Beast of Traal
Let me restate this. If there are zero legal repercussions for those collecting copyrighted material, there is no deterrent on those collecting copyrighted material to make sure the models they train on that material aren't misused. And by misused, I mean used outside the boundaries of Fair Use.Why? That seems like a major leap.
Does it make sense now?
Which is why I didn't refer to it as copyright but as counterfeiting. And if it's close enough that it's considered a derivative work there are other legal issue.What colloquially gets referred to as style cannot currently be copyright except in the most near exact match of cases.
Again, please refer to the real world issue with Studio Ghibli a few months back, that while there currently legal systems have not caught up to AI for widespread style copying, that doesn't mean this isn't something that moving forward shouldn't be protected. The law often lags behind activities-that-should-be-crimes made possible by technology.
Again, the end user doesn't have the training material and can produce copyrighted material without knowing it. The enforcement cannot be on that side, or cannot be solely on that side. There must be provisions in place that if material is collected for Fair Use, it is ONLY used for Fair Use. And that means deterrent needs to affect the collector.Generative ai would be great for that. As long as end users get held accountable for copyrighted materials they produce with an AI then there would be incentive for them to use ai’s that don’t produce copyrighted works, and thus incentive to create such ai’s.
I'm going to assume that since you suggested and are defending it, that you understand what fair use is. Let me remind others:I think it’s worth noting that there’s currently a lot of duplication that qualifies as fair use.
Fair use is a legal doctrine that permits limited use of copyrighted material without permission from the copyright holder, primarily for purposes like criticism, commentary, news reporting, teaching, scholarship, and research.
Can you explain how you see that 'a lot of the duplication' is for the purpose of 'criticism, commentary, news reporting, teaching, scholarship, and research'?
Because I don't see that, especially among generative AI used for art.