Platforms Seed Piracy for AI Training

Authors allege Meta used BitTorrent to download and 'seed' pirated book collections (like Anna's Archive) as part of building or testing LLM datasets, and a federal judge permitted contributory‑infringement claims to be added to the complaint despite criticizing plaintiffs' counsel. If proven, the idea is that platform activity that facilitates peer‑to‑peer distribution of copyrighted works can be framed as direct legal exposure for AI dataset assembly. — If courts accept contributory claims tied to platform torrenting, tech companies may have to change how they acquire, vet, and host training data, with broad effects on AI development and copyright enforcement.