|
|
|
|
|
by danaris
846 days ago
|
|
Copyright has absolutely not "failed". Copyright is still deeply important to prevent behemoths from just straight-up taking stuff individuals wrote and profiting from it with no consequences. For instance, without copyright, traditional publishers could just take everything the authors they currently contract with have written, and every other current author, and publish it without paying the authors a cent. ML training is a legal gray area right now, because it's a new thing, and we haven't had time to properly sit down and both understand what its effects are, and how it should be treated legally. It is possible that this process, when it ends up happening, will be captured by corruption; it is possible that it won't. But using the current frustrations and anger about ML training as evidence that copyright has "failed" is a vast oversimplification that ignores the very real good that copyright does in our society. |
|
It's failing right now to protect millions/billions of people, because we've decided that it's "legal gray area right now".
Maybe it should be, I don't know. I mean maybe it's time we said bye bye to copyright?
There could be flip sides. If the world decides that ML sidesteps copyright then I look forward to the entire corpus of LibGen, SciHub etc being legally released as open models and the overnight demise of Elsevier et al. (I once wrote a fiction about that [0])
My objection here is to seeing the clear wishes of the majority being trodden over roughshod.
[0] https://www.timeshighereducation.com/opinion/2048-informatio...