CEO Liability for AI Training Data

The lawsuit alleges Mark Zuckerberg personally authorized torrenting and massive web scraping to assemble Llama’s training data and instructed teams not to license works. That claim, if proven, would test whether corporate leaders can be held individually responsible for illegal dataset acquisition used to build generative AI models. — If courts treat executive direction as a distinct basis for liability, it could reshape corporate AI strategy, force licensing markets, and make CEO testimony a central feature of AI litigation and regulation.