How will it know what grrm is if it hasn’t read the book or is aware of the content? Pretty sure it does need to read the book in order to generate content similar to the authors style.
Right, but would that be pursued if a human did the same thing? Because there’s a vast amount of fanfiction churned out by human hands, and it’s safe as long as they don’t try to monetize it. Seems like most of the fear is the existential threat that it might suddenly begin writing good stories, and destabilize the entire writing industry as people can just ask for exactly the sort of story they want. For free. (But not actually, because corporations will own all the AI and data).
Well, I mean we kinda are, capitalism and all that. There are thousands of authors of Patreon, Kofi, and the like that you can pay to write you the fanfiction you want. Further, if you don’t know the provenance of a fanfic, how do you tell which ones are the copyright violation? The only way to do so is if you have records of its birth, especially as generative AI improves.
I’m not blind to the plight of creators here, but isn’t the issue that a machine can, in theory, out compete the authors at their own style? If a random human can write Stephen King’s style better than Stephen King, it’s forgiven because that took time, effort, and talent, where a machine doing it alarms us. No author has ever been sued because they read a book and were influenced in their writing, unless they outright plagiarized without attributing. I just think that there needs to be a significant frame shift, since artificially limiting generative AI to protect the current business model instead of allowing it to reshape how people produce and consume media isn’t realistic. The issue is figuring out how creators are still compensated for their work.
People are already building small generative AI projects, so there’s no containing it, and it’s only going to grow.
The thing is you can tell an AI to make a story like grrm and the AI doesn’t even have to read grrm. This is a losing battle.
How will it know what grrm is if it hasn’t read the book or is aware of the content? Pretty sure it does need to read the book in order to generate content similar to the authors style.
Right, but would that be pursued if a human did the same thing? Because there’s a vast amount of fanfiction churned out by human hands, and it’s safe as long as they don’t try to monetize it. Seems like most of the fear is the existential threat that it might suddenly begin writing good stories, and destabilize the entire writing industry as people can just ask for exactly the sort of story they want. For free. (But not actually, because corporations will own all the AI and data).
A human isn’t a product for use by other humans. It’s not the same.
Well, I mean we kinda are, capitalism and all that. There are thousands of authors of Patreon, Kofi, and the like that you can pay to write you the fanfiction you want. Further, if you don’t know the provenance of a fanfic, how do you tell which ones are the copyright violation? The only way to do so is if you have records of its birth, especially as generative AI improves.
I’m not blind to the plight of creators here, but isn’t the issue that a machine can, in theory, out compete the authors at their own style? If a random human can write Stephen King’s style better than Stephen King, it’s forgiven because that took time, effort, and talent, where a machine doing it alarms us. No author has ever been sued because they read a book and were influenced in their writing, unless they outright plagiarized without attributing. I just think that there needs to be a significant frame shift, since artificially limiting generative AI to protect the current business model instead of allowing it to reshape how people produce and consume media isn’t realistic. The issue is figuring out how creators are still compensated for their work.
People are already building small generative AI projects, so there’s no containing it, and it’s only going to grow.