Ugh the ones with the “anti-AI” “licensing”? I honestly don’t know how they expect that to accomplish anything, other than look like a pretentious idiot.
It’s either fair use to train, and the license does nothing, or it is not fair use to train, and the license is extraneous. I explained that to one of them and he said the license isn’t actually supposed to do the “anti-AI” he claims it does. Thus winning the discussion, I guess.
For the US: you already have a copyright on everything you write. Adding CC-BY-NC grants others the right to republish it for non-commercial use, it does not remove any rights at all.
Yes, that is how that works, correct. CC-BY-NC specifically prevents commercial use and as far as I see that, that’s the point of those people that add that footer to their posts.
No, that’s not how it works. Commercial use is already prohibited by default. They are adding no new prohibitions. They are allowing use that is not allowed by default.
I agree it’s not great for public discourse, but you are subject to the whims of the third party’s TOS. You can still publish it, but have to make your own site or pick a third party that won’t sell you out.
Or you just accept that the third party is going to do what it’s going to do. Post your discourse, but don’t expect it to be protected.
That’s not how copyright works. In my example: Publishing code on GitHub still means that code remains under whatever license I put on it, no matter how GitHub feels about that. They can not go ahead and sell my little helper library to Sony for use in their PlayStation OS update to make a buck. If I put it under GPL e.g., Sony would still have to stick to that.
Same goes for posting on a public Lemmy instance. I publish content, I can limit it’s usage. The only thing the instance could do about that is remove my content if their usage of it is in conflict with whatever limitations I put on it.
Should, in some form of future, courts rule that those LLM generating companies can’t just use everything they get their hands on as they please, the poison pill is in their models. That’s why that thing matters now. Could it be done more elegantly than a persistent footer? Certainly. Right now though? I don’t see it.
I agree, that’s how it should work. GitHub as a third party probably had better TOS than most, and would face more uproar if they changed it. But they are not a social media site.
But for regular social media posting, Facebook, Twitter, etc, there is no fighting against their TOS other than abstinence. You can object in writing with one of those footers, but nobody in charge is going to read it or honor it. It will be shoveled into the AI along with everything else. Your only recourse will be an expensive legal fight, and it would be difficult to prove any particular post was in the AI or not. It’s unproven legal ground to say giving a post to an AI for training even qualifies as a “copy” under copyright, or what notification qualifies to exempt your content.
Ugh the ones with the “anti-AI” “licensing”? I honestly don’t know how they expect that to accomplish anything, other than look like a pretentious idiot.
It’s either fair use to train, and the license does nothing, or it is not fair use to train, and the license is extraneous. I explained that to one of them and he said the license isn’t actually supposed to do the “anti-AI” he claims it does. Thus winning the discussion, I guess.
Any better idea how to prevent scraping or, at least, be able to fight it after the fact?
For the US: you already have a copyright on everything you write. Adding CC-BY-NC grants others the right to republish it for non-commercial use, it does not remove any rights at all.
Yes, that is how that works, correct. CC-BY-NC specifically prevents commercial use and as far as I see that, that’s the point of those people that add that footer to their posts.
No, that’s not how it works. Commercial use is already prohibited by default. They are adding no new prohibitions. They are allowing use that is not allowed by default.
Not everywhere is the US. It’s not that simple here.
Yes, that’s why I wrote specifically for the US. Where many prominent AI companies are.
Don’t post on a third party site.
That’s like saying “If you don’t want corporations to steal your sourcecode, just don’t publish it.” That’s just not the right conclusion to come to.
I think it should be absolutely possible to communicate with other people in the open while still maintaining some kind of control over your creation.
I agree it’s not great for public discourse, but you are subject to the whims of the third party’s TOS. You can still publish it, but have to make your own site or pick a third party that won’t sell you out.
Or you just accept that the third party is going to do what it’s going to do. Post your discourse, but don’t expect it to be protected.
That’s not how copyright works. In my example: Publishing code on GitHub still means that code remains under whatever license I put on it, no matter how GitHub feels about that. They can not go ahead and sell my little helper library to Sony for use in their PlayStation OS update to make a buck. If I put it under GPL e.g., Sony would still have to stick to that.
Same goes for posting on a public Lemmy instance. I publish content, I can limit it’s usage. The only thing the instance could do about that is remove my content if their usage of it is in conflict with whatever limitations I put on it.
Should, in some form of future, courts rule that those LLM generating companies can’t just use everything they get their hands on as they please, the poison pill is in their models. That’s why that thing matters now. Could it be done more elegantly than a persistent footer? Certainly. Right now though? I don’t see it.
I agree, that’s how it should work. GitHub as a third party probably had better TOS than most, and would face more uproar if they changed it. But they are not a social media site.
But for regular social media posting, Facebook, Twitter, etc, there is no fighting against their TOS other than abstinence. You can object in writing with one of those footers, but nobody in charge is going to read it or honor it. It will be shoveled into the AI along with everything else. Your only recourse will be an expensive legal fight, and it would be difficult to prove any particular post was in the AI or not. It’s unproven legal ground to say giving a post to an AI for training even qualifies as a “copy” under copyright, or what notification qualifies to exempt your content.
A footnote isn’t going to stop anyone from scraping comments. That’s just reality.