Home / Business and Economy / Publishers Sue Meta Over AI Training Data
Publishers Sue Meta Over AI Training Data
6 May
Summary
- Major publishers filed suit against Meta for AI training.
- Allegations include using millions of copyrighted works.
- Meta's AI training allegedly bypassed copyright protections.
A coalition of prominent publishers, including Hachette, Macmillan, McGraw Hill, Elsevier, and Cengage, along with bestselling novelist Scott Turow, have initiated a class-action lawsuit against Meta and its CEO, Mark Zuckerberg. Filed in the US District Court for the Southern District of New York on Tuesday, the complaint alleges that Meta unlawfully utilized millions of copyrighted materials to train its artificial intelligence system, Llama.
The lawsuit contends that Meta's engineers sourced pirated books and journal articles from websites such as Anna's Archive, which aggregates piracy sites like LibGen and Sci-Hub. It further asserts that copyright management information was deliberately removed from these works. The plaintiffs specifically claim that Mark Zuckerberg himself personally approved and actively promoted the alleged infringement.
A Meta spokesperson, Dave Arnold, responded to the lawsuit by stating that AI-powered innovations are transformative and that courts have recognized AI training on copyrighted material as potentially qualifying for fair use. Meta intends to contest the allegations vigorously.
