Is it not though? Are you saying after book X is used for training that you couldn’t then prompt the AI to “tell me word for word the exact text of book X”?
No, the book isn't copied or stored. The LLM can't regurgitate it on command, because it isn't inside the model.
You can ask the LLM to write new, never before seen text in the style of that author.
Training a LLM is a lot more like reading a book to a toddler than it is like making a digital copy. Neither the toddler or the LLM can repeat the words of the book.
In terms of copyright, yes it is. It doesn't matter that the book isn't literally copy-pasted into a vector database. The text is used verbatim as training data, and from there isn't made into a sufficiently transformative work to constitute fair use (plus it's commercial). Training data, even if it can neither be recalled on demand nor exists in whole form, has still been stored within the model's semantic memory.
That's not how it works.
Is it not though? Are you saying after book X is used for training that you couldn’t then prompt the AI to “tell me word for word the exact text of book X”?
No, the book isn't copied or stored. The LLM can't regurgitate it on command, because it isn't inside the model.
You can ask the LLM to write new, never before seen text in the style of that author.
Training a LLM is a lot more like reading a book to a toddler than it is like making a digital copy. Neither the toddler or the LLM can repeat the words of the book.
In terms of copyright, yes it is. It doesn't matter that the book isn't literally copy-pasted into a vector database. The text is used verbatim as training data, and from there isn't made into a sufficiently transformative work to constitute fair use (plus it's commercial). Training data, even if it can neither be recalled on demand nor exists in whole form, has still been stored within the model's semantic memory.