Meta Accused of Using Pirated Porn to Train Its AI

Meta is facing a new lawsuit from two adult film companies who claim the tech giant illegally downloaded and shared copyrighted porn to train its AI systems.

Filed in a California district court on Friday, the suit comes from Strike 3 Holdings and Counterlife Media. They allege Meta "willfully and intentionally" pirated at least 2,396 adult films since 2018. Strike 3, a major player in the adult entertainment world, is known for aggressively going after copyright violations, especially on torrent platforms.

The lawsuit claims Meta downloaded these videos to train AI systems like its Meta Movie Gen, LLaMA language model, and other tools that rely on large amounts of video content. It also alleges Meta didn’t just download the content without permission—but redistributed it through peer-to-peer networks, speeding up its access to even more material.

Torrenting works by sharing file pieces among users. The suit argues Meta participated in this process, meaning it didn’t just consume the pirated files, it helped circulate them too—possibly gaining technical or data benefits along the way.

Strike 3 says its websites draw over 25 million visitors monthly and is asking for damages that could hit $359 million. The company also wants a court order banning Meta from ever using its content again.

This isn't Meta’s first brush with AI-related copyright cases. Authors like Sarah Silverman and Ta-Nehisi Coates have also sued Meta for allegedly using their work without consent. That case was recently dismissed. Still, these lawsuits highlight a growing tension between protecting creators' rights and pushing ahead with AI development.