Amy | Beauty and Books

For those of you unaware, it’s been recently unearthed by the @theatlantic that @meta used a digital library of pirated books to train their Large Language Model, Llama 3. This library consisted of over 80tb of literary works, constituting the theft of millions of books, likely from every author you’ve ever heard of, including every novel I’ve ever written. None of these authors were compensated for the use of their work, nor asked permission for its use. None of us. This is not simply a matter of theft - although theft at this scale at the behest of a billionaire is particularly egregious. I’m not being hyperbolic when I say that LLM (aka AI) technology represents a clear and present danger to the human artistic endeavor. Despite claims to the contrary, it is not a tool through which people can unlock their hidden artistic gifts. It is a plagiarism machine, trained on the stolen works of hundreds of thousands of human artists, developed solely to supplant the very artists from which it has stolen. There are several lawsuits already underway against Meta over this theft. A full list of the stolen work is available through @theatlantic . If you want to educate yourself further about the dangers of LLM/AI tech (as opposed to doomscrolling the gibberish of armchair experts on Reddit), I suggest you start with the work of @tristanharris @humanetech and their brilliant podcast, Your Undivided Attention.