Reports says that Meta has refuted claims that its artificial intelligence (AI) model Llama was developed using content that was protected by copyright from well-known novels. On September 18, Meta asked for the accusations brought by author Sarah Silverman and other authors accusing the business of breaking copyrights to train its AI system to be dismissed in a San Francisco federal court. The parent corporation of Facebook and Instagram claimed that using materials to train its systems constituted “fair use” and was “transformative.”
Meta compared their situation to Authors Guild v. Google, Inc., where fair use was established for Google’s unauthorised use of books to construct an internet search engine. The business advocated for debating the “core issue” of copyright fair use “another day, on a more fulsome record” The plaintiffs were unable to give particular information or outputs linked to their content, according to Meta as well.
Attorneys for the writers expressed confidence in their allegations and promised to continue with “discovery and trial.” OpenAI made an attempt to deny some of the assertions in August using the same defences as Meta.
One of numerous lawsuits against Big Tech corporations over copyright and data infringement with the rise of AI, the initial case against Meta and OpenAI was filed in July. Unidentified engineers filed a class-action lawsuit on September 5 against OpenAI and Microsoft, claiming that the businesses utilised scraping techniques to gather sensitive data while developing their AI models. After amending its privacy statement, Google was hit with a similar complaint in July. The company was accused of using copious amounts of data—including content protected by copyright—in its AI training.