The problem this argument has is that it shifts copyright from being judged on the content seeking copyright to the process used to create it. If an author reads 100 books and writes a book off that research the output may or may not breech copyright depending on the final content.
good point. If that author reads 100 books for research do they not have to pay for those books or get permission to source? I think they do. As you say, there is a difference between fiction/non fiction but the easiest way to quantify impact is to pay for the book, like buying a cinema ticket is
Not if they, for instance, read them in a library, or if the books are freely available online- and much of the information people are worried about AI training on is freely available
True enough but a text like that would be expected to contain references and a bibliography, would it not? Nobody touting the glorious future of gen AI is even pretending to address that bit.
I'm thinking of works of fiction here. For example if every fantasy author since Tolkien was held to this copyright standard then Christopher Tolkien would have been much, much busier
Fair point. Maybe the key difference is the industrialisation and automation. Rather than a considered osmosis and distillation into a new work, you press a button and derivative crap comes out in great volume.
Right, but this takes it into entirely new territory, and people need to recognise that
You think they don't? It seems to me the unfairness of being able to effortlessly monetise others' works thru AI is rightly upsetting a lot of creators. The language around theft etc stems from this and the lack of representation of creators in the drive for wider use of gen AI
I don't think it's "effortlessly" that's the problem. Creating AI takes an awful lot of effort- the problem is a combination of the consequences of AI and the moral worth people attach to creative arts Vs technology
Ok I see. Starting to stray beyond my knowledge and expertise so will leave it there. Nice chatting