3 ms·
In the US, everything you write anywhere online is copyrighted by you, unless you sign a copyright assignment agreement. It's automatic any time you put an expr
by JohnFen 3y ago
In the US, everything you write anywhere online is copyrighted by you, unless you sign a copyright assignment agreement. It's automatic any time you put an expression into a fixed form, and there is no way to revoke that copyright.
- nonrandomstring 3y agoAs I understand it copyright has failed. Or rather we are into an age of naked double standards where courts will enforce the copyright of big-tech against you for "stealing" a movie, but will not enforce your copyright against big-tech for "stealing" your data for its AI. EDIT: some links [0,1] [0] https://www.reuters.com/legal/litigation/us-judge-finds-flaws-artists-lawsuit-against-ai-companies-2023-07-19/ https://www.reuters.com/legal/litigation/us-judge-finds-flaw... [1] https://www.hollywoodreporter.com/business/business-news/artists-copyright-infringement-case-ai-art-generators-1235632929/ https://www.hollywoodreporter.com/business/business-news/art...
- danaris 3y agoCopyright 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.
- nonrandomstring 3y agoFair comments. Please allow me to rephrase that. 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-information-unchained-and-even-schoolgirls-win-nobel-prizes https://www.timeshighereducation.com/opinion/2048-informatio...
- danaris 3y ago> I mean maybe it's time we said bye bye to copyright? This is exactly the kind of oversimplified, baby-with-the-bathwater proposal I was talking about. No, we should not "say bye bye to copyright". We should actually take the harder, more complex steps, requiring actual critical thinking and analysis, to fix the problem, rather than just pretending that a one-step grand gesture will be a magic bullet.
- nonrandomstring 3y ago> We should actually take the harder, more complex steps, requiring actual critical thinking and analysis, Those are fine words. We're all about critical thinking and analysis round here. But way I see it, folks already did some real hard critical thinking and their analysis was "bollocks to that!" And the judges said, "sorry the law that applies to you doesn't apply when big money is involved". One rule for you, another rule for them. So I'm kinda thinking we'll maybe have to get a little more critical than you might be comfortable with. > to fix the problem It's always a good idea to pause right there. What is the problem? I mean seriously... what exactly is the problem going on here? Because from where I see it, the problem is a massive power imbalance And it's a structural one. Because AI training compute and global crawling/scraping is expensive and in the hands of the few. I don't think this problem would look the same if every kid was running AI training on a Raspberry Pi, and hoovering down JSTOR like Aaron Swartz. People would be getting arrested, no?