It is missing one point: as a creator, I want to be able to forbid you from training on my creations. And the only tool that could enable that is the copyright enforcement over AI training.
Exactly
If there was an opt out system that was actually respected then this wouldn’t be a problem. But as it stands, artists have no control over if their work is used for NN training.
I don’t want my work used to train models, which should be a completely valid stance to have. Open Source or not really doesn’t matter in the grand scheme of it.
The AI companies shown that they are incapable of regulating themselves on this topic, and so people with art at stake should force their hand.
Open source or not doesn't matter here, what matters is the copyright. If even Disney can defend works they own (whatever their ethics), so should anyone else.
100% agreement from me again. Non-artists don’t have anything at stake, so they’re perfectly happy with the established copyright rules are demolished. People keep countering with the open source idea, which completely misses the entire point of our arguments. A model being open source does not excuse the stealing of training data.
IMO individual copyright should be strengthened and corporate copyright weakened, but that’d be next to impossible to pass.
Too bad. You can "forbid" all you want. Don't mean shit. Vote for much stronger laws. By much stronger I mean no pay a fine and continue. I mean jail.
No. I reject you claiming such a power to deny.
That's exactly what's at stake, waiting to be sufficiently litigated. And I hope that creators will win, and that they would be able to tell if they allow richest big tech companies in the world to train on their creations.
Likewise, I hope they don't win, as that will give the richest tech companies so much more of a stranglehold.
I doubt there's any chance of it happening anyway, since there's a ton of money to be made and and there's already countries which have rules this will never happen (Like Japan ), so it would mean they become the AI powerhouses
They have already trained on those creations though. Including the newer stuff just released today. How will you claw that back?
If you do stuff, earn from it, and ignore parties and their rights, you are forced to compensate. I guess it will be peanuts though.
They could shut down the previous models that were trained on invalid works. Sucks to suck but that's what you get when you do everything in your power to skirt the law.
Yeah, and the same thing would happen if e.g. PII or HIPAA related would end up in trained model. The fact that some PII or health data ended up being publicly available, doesn't mean that automatically you can process or store such data, and train on such data.
This has already been proven by google security researchers who got several of the big "AI" bots to spit out copyrighted materials and PII from their training data sets which the "AI" creators claimed was not stored.
It's not stored as the full material though. If a human that can sing a copyrighted song is not considered to have a recording of the copyrighted song in their brain, so too are LLMs able to spit out their training data without having to store them.
How do you know what it's storing? I certainly don't, but I know what the security researchers have found that proved it was storing copyrighted material and real people's private info or PII.
You being able to spit people's name and personal details doesn't mean you are keeping a database of those details in your brain. It's all just neurons and the connection between them that can be triggered to extract those details out.
LLMs also attempt to mimic this method of not storing direct information, but tweaking parameters to 'learn' the information. Inside LLMs are just a bunch of parameters that if not well-designed, can be made to spit out what they have learnt. That doesn't mean they store those information as is.
That's what they tell you about it I'm sure, but what proof do you have?
It's not just what they tell you. There are plenty of publicly accessible LLM models. Go and download them and open the files up. Surely if they are storing these things as complete data, you can easily find them by poking around the files instead of having to make then spit it out.
lol, if you want that, keep your pictures for you, else you had to forbid every human to look at your pictures and they could resemble your style
And I want a law making you pay me 500$ for reading your posts.
Copyright law already extends beyond what society finds reasonable. It's routinely broken by normal people without them even thinking about it. It's even broken by those vested in it both corporations and individual artists.
Finally you are not getting the copyright law you want ( nor should you, you a minority, a special interest ), big corps are. They might be 'content' corps or tech or both but they certainly won't make a law to benefit either society as a whole or you as a small artist.
It is missing one point: as a creator, I want to be able to forbid you from training on my creations. And the only tool that could enable that is the copyright enforcement over AI training.
Exactly
If there was an opt out system that was actually respected then this wouldn’t be a problem. But as it stands, artists have no control over if their work is used for NN training.
I don’t want my work used to train models, which should be a completely valid stance to have. Open Source or not really doesn’t matter in the grand scheme of it.
The AI companies shown that they are incapable of regulating themselves on this topic, and so people with art at stake should force their hand.
Open source or not doesn't matter here, what matters is the copyright. If even Disney can defend works they own (whatever their ethics), so should anyone else.
100% agreement from me again. Non-artists don’t have anything at stake, so they’re perfectly happy with the established copyright rules are demolished. People keep countering with the open source idea, which completely misses the entire point of our arguments. A model being open source does not excuse the stealing of training data.
IMO individual copyright should be strengthened and corporate copyright weakened, but that’d be next to impossible to pass.
Too bad. You can "forbid" all you want. Don't mean shit. Vote for much stronger laws. By much stronger I mean no pay a fine and continue. I mean jail.
No. I reject you claiming such a power to deny.
That's exactly what's at stake, waiting to be sufficiently litigated. And I hope that creators will win, and that they would be able to tell if they allow richest big tech companies in the world to train on their creations.
Likewise, I hope they don't win, as that will give the richest tech companies so much more of a stranglehold.
I doubt there's any chance of it happening anyway, since there's a ton of money to be made and and there's already countries which have rules this will never happen (Like Japan ), so it would mean they become the AI powerhouses
They have already trained on those creations though. Including the newer stuff just released today. How will you claw that back?
If you do stuff, earn from it, and ignore parties and their rights, you are forced to compensate. I guess it will be peanuts though.
They could shut down the previous models that were trained on invalid works. Sucks to suck but that's what you get when you do everything in your power to skirt the law.
Yeah, and the same thing would happen if e.g. PII or HIPAA related would end up in trained model. The fact that some PII or health data ended up being publicly available, doesn't mean that automatically you can process or store such data, and train on such data.
This has already been proven by google security researchers who got several of the big "AI" bots to spit out copyrighted materials and PII from their training data sets which the "AI" creators claimed was not stored.
It's not stored as the full material though. If a human that can sing a copyrighted song is not considered to have a recording of the copyrighted song in their brain, so too are LLMs able to spit out their training data without having to store them.
How do you know what it's storing? I certainly don't, but I know what the security researchers have found that proved it was storing copyrighted material and real people's private info or PII.
You being able to spit people's name and personal details doesn't mean you are keeping a database of those details in your brain. It's all just neurons and the connection between them that can be triggered to extract those details out.
LLMs also attempt to mimic this method of not storing direct information, but tweaking parameters to 'learn' the information. Inside LLMs are just a bunch of parameters that if not well-designed, can be made to spit out what they have learnt. That doesn't mean they store those information as is.
That's what they tell you about it I'm sure, but what proof do you have?
It's not just what they tell you. There are plenty of publicly accessible LLM models. Go and download them and open the files up. Surely if they are storing these things as complete data, you can easily find them by poking around the files instead of having to make then spit it out.
lol, if you want that, keep your pictures for you, else you had to forbid every human to look at your pictures and they could resemble your style
And I want a law making you pay me 500$ for reading your posts.
Copyright law already extends beyond what society finds reasonable. It's routinely broken by normal people without them even thinking about it. It's even broken by those vested in it both corporations and individual artists.
Finally you are not getting the copyright law you want ( nor should you, you a minority, a special interest ), big corps are. They might be 'content' corps or tech or both but they certainly won't make a law to benefit either society as a whole or you as a small artist.