ChatGPT and other chatbots were trained using protected works without permission or compensation. This is definitely one of the reasons it is as good at doing what it does. In the literary world, many books that had been pirated on bittorrent and other sites were consumed...despite the fact that researchers knew full well that most torrent type sites are rife with pirated material. A number of sources have been identified and confirmed independently. Books3 is one of the primary ones.
183,000 books have been identified so far as being used in training data.If you're a writer, check out the links below and know that the Authors Guild has your back. Check out the links below.
