In a copyright infringement lawsuit against Meta Platforms, new filings reveal that Meta, despite legal warnings, used pirated books to train its AI models. The consolidated complaint includes chat logs suggesting that Meta was aware of potential copyright issues. The logs feature a researcher discussing legal concerns with Meta's legal department, acknowledging that the dataset used for training, known as "ThePile," could not be used for legal reasons. This lawsuit, joined by authors like Sarah Silverman and Michael Chabon, reflects a broader trend of content creators suing tech companies over alleged unauthorized use of copyrighted works for generative AI models. If successful, these cases may impact the generative AI industry and raise costs for companies using copyrighted content without compensation. The filing also comes amid growing legal scrutiny, with new European rules potentially requiring AI companies to disclose data used for training.