Autores demandan a Microsoft, Meta y Bloomberg sobre IA y los derechos de autor
A group of authors has filed a federal lawsuit accusing major technology companies, including Meta, Microsoft, and Bloomberg, of using copyrighted works without permission to train generative AI models. The plaintiffs allege that these corporations exploited a controversial dataset containing thousands of pirated books, arguing that such actions constitute theft rather than legitimate innovation, despite the companies’ claims of protection under fair use doctrines. This legal action highlights the growing tension between AI developers and content creators regarding data sourcing. The dispute centers on the validity of scraping extensive digital libraries to create large language models. It raises critical questions about intellectual property rights in the age of artificial intelligence, challenging the assumption that training AI on human creativity is automatically permissible without explicit consent or compensation. This case is profoundly relevant to open data because it directly impacts the availability of high-quality, legally safe datasets for research and development. If courts rule that such scraping violates copyright, the landscape of open AI training data could shrink significantly. Consequently, the open data community must advocate for clearer legal frameworks that balance innovation with authors’ rights, ensuring sustainable and ethical data practices for future technological advancements.
Source: larepublica.coPublished on 2023-10-19