Meta's Llama AI model surprisingly memorized a large portion of J.K. Rowling's Harry Potter, raising legal questions.
The study shows Llama accurately reproduced over 40% of the book, highlighting potential copyright infringement issues.
This discovery significantly impacts the ongoing debate about fair use in AI training data and the legal ramifications.
The case underscores the challenges of balancing technological advancement with the protection of intellectual property rights.
