A legal expert revealed that Meta's AI model, LLaMA, can reproduce copyrighted content, such as entire passages from books, including the Harry Potter series. This poses a significant legal liability for Meta, as it undermines the claim that AI produces transformative and original content. Unlike other AI systems, LLaMA's behavior resembles that of a repository for copyrighted work. The situation raises critical questions about intellectual property rights in relation to AI training methods and the implications for companies like Meta and their leadership.
Meta's LLaMA AI has the alarming ability to reproduce copyrighted texts verbatim, posing significant legal risks for the company and its CEO.
The study led by Stanford's Mark Lemley indicates that the way LLaMA processes training data may lead it to act as a repository of copyrighted material instead of creating new, transformative content.
Collection
[
|
...
]