George R.R. Martin and other authors sue OpenAI for copyright infringement
George R.R. Martin and other authors sue OpenAI for copyright infringement
Martin and other popular authors said OpenAI stole their copyrighted work.
I mean this isn't miles away from what the writer's strike is about. Certainly I think the technology is great but after the last round of tech companies turning out to be fuckwits (Facebook, Google etc) it's only natural that people are going to want to make sure this stuff is properly regulated and run fairly (not at the expense of human creatives).
If the models trained on pirated works were available as a non-profit sort of setup with any commercial application being banned I think that would be fine.
Business owners salivating over the idea that they can just pocket the money writers and artists would make is not exactly a good use of tech.
Copyright law in general is out of date and needs updating it's not just AI that's the problem that's just the biggest new thing. But content creators of traditional media have been railing against what they perceive as copyright violation for ages.
Look at Nintendo and Let's Plays.
The law is the problem here. Not AI.
I've expressed my opinions on this before which wasn't popular, but I think this case is going to get thrown out. Authors Guild, Inc. v. Google, Inc. has established the precedent that digitalization of copyrighted work is considered fair use, and finetuning an LLM even more so, because LLMs ultimately can be thought of as storing text data in a very, very lossy comprssion algorithm, and you can't exactly copyright JPEG noise.
And I don't think many of the writers or studio people actually tried to use ChatGPT to do creative writing, and so they think it magically outputs perfect scripts just by telling it to write a script: the reality is if you give it a simple prompt, it generates the blandest, most uninspired, badly paced textural garbage imaginable (LLMs are also really terrible at jokes), and you have to spend so much time prompt engineering just to get it to write something passable that it's often easier just to write it yourself.
So, the current law on it is fine I think, that pure AI generated contents are uncopyright-able, and NOBODY can make money off them without significant human contribution.
To be the devil's advocate (or GRRM's attorney), I see the merit of his and other authors' concerns. Chat GPT makes it feasible to generate short stories in their world and with their characters, which can easily replace their licensed products. This is not just their main work, but also other products that generates them some revenue stream.
Example: A friend of mine is using Chat GPT to generate short bedtime stories for his daughters. A typical request is something like this: "Please write a five paragraph story where Elsa from Frozen meets Santa Claus. Together, they fly in Santa's sleigh over the world, and Elsa is magicking snow on all Christmas trees." Normally, you'd buy a Disney-licensed book of short Christmas stories (I have one for my kids), but Chat GPT is more flexible and free.
Same goes for GRRM. He doesn't write Children stories, but one can still prompt Chat GPT to produce stories from the universe, which scratch the ASOIAF itch. This substitutes the officially licensed products and deprives the author of additional revenue stream. Just for the fun of it, I prompted Chat GPT: "Hello GPT-3.5. Please write a four paragraph story set in the Game of Thrones universe. In this story, Jon Snow and Tyrion Lannister go fishing and catch a monster alligator, which then attacks them." It produces a surprisingly readable story, and if I were a fan of this universe, I can imagine myself spending a lot of time with different prompts and then reading the results.
(On a side note,AI-generated content already has at least one group of victims: the authors of short fiction. Magazines like Clarkesworld were forced to close submissions of new stories, as they became overwhelmed with AI-generated content.)
"LLMs allow anyone to generate — automatically and freely (or very cheaply) — text that they would otherwise pay writers to create" My heart bleeds for them 🙄
That new technology is going to make it harder for us to earn income. As if automation and other improvements over the years hasn't diminished other positions and they should somehow be protected at the cost of improvements for everyone as a whole
Julia was twenty-six years old... and she worked, as he had guessed, on the novel-writing machines in the Fiction Department. She enjoyed her work, which consisted chiefly in running and servicing a powerful but tricky electric motor... She could describe the whole process of composing a novel, from the general directive issued by the Planning Committee down to the final touching-up by the Rewrite Squad. But she was not interested in the final product. She "didn't much care for reading," she said. Books were just a commodity that had to be produced, like jam or bootlaces.
I don't want AI trained on pirated works 'found' by scrubbing the entire internet.
This is the best summary I could come up with:
According to the complaint, OpenAI “copied plaintiffs’ works wholesale, without permission or consideration” and fed the copyrighted materials into large language models.
The authors added that OpenAI’s LLMs could result in derivative work “that is based on, mimics, summarizes, or paraphrases” their books, which could harm their market.
OpenAI, the complaint said, could have trained GPT on works in the public domain instead of pulling in copyrighted material without paying a licensing fee.
This is the latest lawsuit against OpenAI from popular authors — Martin wrote Game of Thrones, Grisham’s many books have been turned into films, and so on — alleging copyright infringement.
Amazing Adventures of Kavalier and Clay writer Michael Chabon and others sued the company for using their books to train GPT earlier in September.
Comedian Sarah Silverman and authors Christopher Golden and Richard Kadrey also sought legal action against OpenAI and Meta, while Paul Tremblay and Mona Awad filed their complaint in June.
The original article contains 323 words, the summary contains 157 words. Saved 51%. I'm a bot and I'm open source!
Shit that's actually not the worst idea - get a team of these AI to finish TWoW and ADoS.
Guaranteed to be better than what we've got so far.
So basically he feels threatened by an extrapolator? In my opinion that's something a self-respecting author wouldn't do.
Boomers who don't understand technology yelling into the wind.
Everything is derivative. Adapt and overcome, instead of gatekeep.