Meta received a major legal decision in the AI copyright case brought by 13 authors, alleging that the company had illegally trained its AI system on its work without permission. On Wednesday, Judge Vince Chbiaya ruled in favor of Meta, saying that “according to this claim, a summary decision on defense of its fair use is entitled that copying these plaintiffs’ books as LLM training data is a violation.”
However, the judge also identified some of the weak points in the Big Tech AI’s efforts and meta arguments, in which his actions were defended as a fair use. Judge Chibariya said, “This decision is not on the suggestion that the use of copyright content for training of meta -language models is halal.”
“It’s just for the suggestion that these plaintiffs made false arguments and failed to prepare records in support of the right.” The decision pursued the victory of the major use of anthropic from a separate federal judge yesterday, which ruled that it was a fair use to train their models on legally purchased copies of books.
Judge Chbia says two arguments about the two fair use of the authors “clearly losing:” Meta’s Lama AI’s abilities to reproduce the text of the text from her books, and they use their works to train their AI models without any permission to reduce their tasks. The judge wrote, “Lama is not able to produce a lot of text from the plaintiffs’ books, and the plaintiff is not entitled to the market to license his work as a training data.”
According to Judge Chbia, the plaintiffs did not work enough for a “potentially winning argument” that a copy of the meta would generate a product that would potentially flood the market, which would reduce the market. ” He also discussed the anthropic decision, saying that Judge William al -Soph aside the concerns about the generation of the generation, AI “can be taken to the market for works that are trained.”
