苹果被指控用盗版书训练AI，遭集体诉讼

Core Viewpoint - A collective lawsuit has been filed against Apple by two academic authors, alleging that the company's AI system, "Apple Intelligence," was trained using a database containing pirated works without authorization [2] Group 1: Lawsuit Details - The plaintiffs claim that Apple used pirated content from the "Books3" shadow library to train its AI, specifically referencing their works "Illusion Champions" and "The Trick of Thinking" [2] - Evidence indicates that Apple acknowledged using "The Pile" dataset, which included over 180,000 books from "Books3," during the training of the OpenELM model in April 2024, with the plaintiffs' works included [2] - Although "Books3" was removed in October 2023 due to copyright disputes, the plaintiffs assert that Apple had already completed the copying and training process [2] Group 2: Legal Challenges - The case faces multiple challenges, including Apple's lack of disclosure regarding the training documentation, making it difficult to provide evidence [2] - Previous rulings have generally found that obtaining individual authorizations for vast amounts of data is not feasible [2] - A prior case involving Anthropic ruled that training constituted "fair use" but acknowledged storage infringement, adding complexity to this case [2] Group 3: Plaintiffs' Demands - The plaintiffs are seeking a jury trial, monetary compensation, and an injunction against the sale of the infringing results [2] - Under U.S. copyright law, if willful infringement is established, Apple could face up to $150,000 (approximately 1.07 million RMB) per work [2] - As of now, Apple has not responded to the lawsuit [2]