Datasets:
People going bananas
Damn, I haven't seen this many people going bonkers in quite a while.
We don't like our work being stolen
I get it, but on the other hand once it's available online, your information is no longer your own. Besides, these types of datasets are used to train models on how get better at creating worlds, roleplaying and much more. Is it that bad that people can use this to bring characters to life within a virtual world similar to the stories written in AO3? No one is monetizing off of this, as these models are open source and available to people to use as they wish.
I get it, but on the other hand once it's available online, your information is no longer your own. Besides, these types of datasets are used to train models on how get better at creating worlds, roleplaying and much more. Is it that bad that people can use this to bring characters to life within a virtual world similar to the stories written in AO3? No one is monetizing off of this, as these models are open source and available to people to use as they wish.
- Just because something is publicly available for consumption, doesn’t automatically mean it’s there for the taking, so to speak. And there are clear laws outlining that. What I’ve seen get thrown around here a lot is, that fanfiction in itself is copyright infringement in itself, which it’s not.
Fanfiction falls under Fair Use, which protects it. The ideas and storylines are by the fanfiction author giving said author all rights to their story. And aside from fanfiction, if we want to skirt away from the issue of fanfiction not being viewed as belonging to authors, there are tons of original stories up on the site. Taking those without the explicit consent and permission is, let’s call it what it is Theft and Stealing from authors.
Touching on the topic of a machine spitting out amalgamations of people’s hard work, be it fanfiction or original works, who have never been informed of said theft, is plain wrong. Even if the idea might be to make it open source and available for free and everyone to “enjoy”.
If AI can create something like the stolen stories, why not skip that part and simply read the original on the original site?
At the end of the day, it’s not about the issue of material being monetized, but about the way said material was obtained illegally. Which, I seem to remember, is not allowed on this site per its Terms of Use.
I hope that answers your question and have a good day, even if this comment is just me screaming into the void.
Touching on the topic of a machine spitting out amalgamations of people’s hard work, be it fanfiction or original works, who have never been informed of said theft, is plain wrong. Even if the idea might be to make it open source and available for free and for everyone to “enjoy”.
If AI can create something like the stolen stories, why not skip that part and simply read the original on the original site?
If George R.R. Martin created a series of books, why do people create fan fiction in that universe and not just read the books? Why do you watch the movies based on comic books and fiction books? Why not just read the original and be happy with it? Why do people say "well, they took the franchise in a weird way. I would have done it differently.". If I go to Ao3, how many fan fics in the some established universe will I find?
My point here is that people want to get lost and enjoy stories. Maybe they read something, liked it and wanted to create characters in that world. Or maybe just see if they can continue the story in their own way. I get it that you want to be credited for your work and want to be asked if you are ok with having your fiction used as part of training. However, you are barking up the wrong tree. You are shouting at an individual who is offering this data for free for the open source community to calibrate their models. Your stories and data has been sold millions times over to big companies to train these models to begin with, but those companies have no issue with it since they can do it under the table, hence closed source. Why not shout at Meta who used terabytes of pirated material to train their models?
If George R.R. Martin created a series of books, why do people create fan fiction in that universe and not just read the books? Why do you watch the movies based on comic books and fiction books? Why not just read the original and be happy with it? Why do people say "well, they took the franchise in a weird way. I would have done it differently.". If I go to Ao3, how many fan fics in the some established universe will I find?
My point here is that people want to get lost and enjoy stories. Maybe they read something, liked it and wanted to create characters in that world. Or maybe just see if they can continue the story in their own way. I get it that you want to be credited for your work and want to be asked if you are ok with having your fiction used as part of training. However, you are barking up the wrong tree. You are shouting at an individual who is offering this data for free for the open source community to calibrate their models. Your stories and data has been sold millions times over to big companies to train these models to begin with, but those companies have no issue with it since they can do it under the table, hence closed source. Why not shout at Meta who used terabytes of pirated material to train their models?
See, if we were George R.R. Martin, this wouldn’t even be an argument. But my whole point, and you said it too, is that not crediting authors is wrong. There are people willing to train Open Source AI models on their text, but that route was not taken. The route chose instead, was taking from a vulnerable source and millions of authors not capable of defending themselves in a way an established household name like George R.R. Martin would be.
The root of the “problem” is the same. Insufficient material, whether it be text or video, and fans that fill that space with their own fanworks . But that, and I know I’m beating a dead horse here, is not the issue at hand. The issue is millions of stolen stories from a Non-Profit Organization and an open disrespect toward those who have created them. Yes I know information is being sold like it’s nothing but that doesn’t mean that I am barking up the wrong tree.
I believe most people speaking up at the moment are being more than polite, while hours and hours of hard work are just being scraped from them.
And let’s not bring giants like Meta into this discussion. There have been widesweeping protests and outcry. I condemn those who are being rude and threatening even at the moment, that is not doing anyone any good.
However, I think it’s more than reasonable to attempt to do something while you can. And someone who has blatantly stolen, cannot cry about getting backlash. Threats, no. But critic of their actions is absolutely fine.
If you feel like I was shouting at you, that wasn’t my intention.
Before speaking up (or in this case, barking at trees), have you checked that the dataset actually contains the alledged works that you claim were "stolen"? Or are you running with a narrative created by people who clearly lied to you, and that you believe blindly, instead of thinking with you own head and reading what the owner of the dataset actually says? The dataset doesn't contain images, therefore nothing was stolen from you. You fell for blind and pointless moral panic.
Before speaking up (or in this case, barking at trees), have you checked that the dataset actually contains the alledged works that you claim were "stolen"? Or are you running with a narrative created by people who clearly lied to you, and that you believe blindly, instead of thinking with you own head and reading what the owner of the dataset actually says? The dataset doesn't contain images, therefore nothing was stolen from you. You fell for blind and pointless moral panic.
"The dataset was created by processing works with IDs from 1 to 63,200,000 that are publicly accessible."
If we have publicly accessible works, they are in there. So yes. We do know.
"I get it, but on the other hand once it's available online, your information is no longer your own. Besides, these types of datasets are used to train models on how get better at creating worlds, roleplaying and much more. Is it that bad that people can use this to bring characters to life within a virtual world similar to the stories written in AO3? No one is monetizing off of this, as these models are open source and available to people to use as they wish. "
::LOUD BUZZER SOUNDS::
INCORRECT!!
Just because it's posted on the internet means that anyone can read it; it's been proven many, MANY times that copyright law (both US and international) apply to material posted online. Your argument is invalid on it's face.
It IS bad that people would do this without my consent. I did not, and NEVER WILL consent to ANY of content of my original works to be fed into an AI or monetized without proper licensing agreements in place.
NONE of my works in AO3 are released under ANY form of Open Source License unless I've explicitly stated such. I object to my content being used to train ML systems without my consent- it offends me on every level I can think of. Go ask Disney if you can feed stories of their primary mascot into a ML system without their consent, and I expect you'll get a strongly worded letter back from their lawyers.