Federal judge sides with Meta in lawsuit over training AI models on copyrighted books

A federal judge stood alongside Mita on Wednesday in a lawsuit against the company by 13 books of the book, including Sarah Silverman, who claimed that the company had trained illegally the artificial intelligence models for their copyright works.

Federal judge Vince Chapria issued a brief ruling – which means that the judge was able to make a decision on the case without sending it to the jury – in favor of Meta, and found that the company’s training in artificial intelligence models on copyright books in this case decreased according to the “fair use” doctrine of the law of copyright and thus was legal.

The decision comes just a few days after a federal judge alongside Anthropor in a similar suit. Together, these cases are formed to be a victory for the technology industry, which has spent years in legal battles with media companies that argue that trading artificial intelligence models on copyright works is fair use.

However, these decisions are not a comprehensive victory for some companies that were hoping for that – both judges note that their cases were limited in the range.

Judge Chapria explained that this decision does not mean that all the training of the artificial intelligence model on the work of copyright is legal, but that the plaintiffs in this case “presented the wrong arguments” and failed to develop sufficient evidence to support the correct security.

Judge Chapria said in his decision: “This ruling does not fall into the proposal that the use of dead materials protected by copyrights to train its language models is legal.” Later, he said: “In cases that involve uses such as Meta, the prosecutors seem to win, at least when these cases have better sophisticated records on the market effects of the defendant’s use.”

Judge Chapria spent that the use of Meta for copyrights in this case was transformed – which means that the company’s AI models not only reproduced the authors’ books.

Moreover, the prosecutors failed to persuade the judge that a dead copy of the books has harmed the market for these authors, a major factor in determining whether the copyright law has been violated.

“Prosecutors have not provided any meaningful evidence to reduce the market at all.”

Anthropor and Meta victories include models of artificial intelligence on books, but there are many other active lawsuits against technology companies to train artificial intelligence models on other copyright works. For example, the New York Times sues Openai and Microsoft to train artificial intelligence models on news articles, while Disney and Universal sued Midjourney to train artificial intelligence models on movies and TV shows.

Judge Chapria noticed in his decision that the defenses of just use depend greatly on the details of the case, and some industries may have a fair use argument than others.

“It seems that the markets for certain types of businesses (such as news articles) may be more vulnerable to indirect competition from artificial intelligence outputs,” Chapria said.