NEW YORK (AP) — John Grisham, Jodi Picoult and George R.R. Martin are among 17 authors suing OpenAI for “systematic theft on a mass scale,” the latest in a wave of legal action by writers concerned that artificial intelligence programs are using their copyrighted works without permission.
I respectively disagree there. If a writer didn’t give any concent whatsoever to give an A.I. a copy of their written works for an A.I. to train or base anything on. I think it’s a fair case of theft. Theoretically, someone would be able ask an A.I. to recite an entire book for them. Without ever having to pay its author any contribution.
Did the writer give me permission to read their book, which I used to learn to write better and sell those works?
Did Michelangelo give every art student that learned from his works permission to learn from his work and then produce works in a similar style on their own to sell for profit?
The thing is that humans are doing something drastically different from the current generation of AIs. If you tell Dall-e to come up with something influenced by El-Greco Cezanne and Toulouse-Lautrec you always get a generic impressionist painting you don’t get Picasso. The "AI"s we have are mapping and transformation tools with a random number generator attached. In a way they are more akin to a DJ remixing a music (which requires permission from the copyright owner) than a musician creating their own song.
You mean that it’s far more like a human then, since there were hundreds of contemporaries of Picasso who didn’t create cubism?
(I think you’re arguing from an ethical standpoint whereas OP was arguing legally, but anyway…)
No, that shouldn’t happen. If an AI were ever able to recite back its training data verbatim, that AI would be overfitting. It happens by accident sometimes early on in development when your training data is too small and your model is too big, but it’s an error, and is something to be avoided and corrected.
The whole point of training is to get it to a point where it can’t recite back any of its training data. In order for that to happen, the AI is forced to sort of generalize and abstract (sorry for anthropomorphizing) its training data. That’s the only way to get it to be able to generate something new, which is the whole point of the endeavour.
Long story short, if an AI could recite back an entire book, by definition it could not be an AI, and it wouldn’t resemble any of the popular LLMs we have now like ChatGPT. (But you may see snippets and pastiches and watermarks show up)