For years, Meta has allegedly downloaded copyrighted adult content via torrent networks to train its AI models. This is detailed in a new lawsuit filed by two companies from the adult industry.
Strike 3 Holdings and Counterlife Media have accused Mark Zuckerberg’s firm of “wilful and deliberate” copyright infringement on at least 2,396 films since 2018.
According to the document, Meta downloaded adult films “to obtain content for training its Movie Gen film generation model and other neural networks that use video as training material.”
The company allegedly did this without permission and with the intent to distribute the videos “to accelerate the download of vast amounts of other content.”
Popular torrent networks are used not only for downloading media files. They are peer-to-peer data exchange platforms where users directly share file fragments with each other.
The lawsuit claims that Meta not only downloaded pornographic content but also provided it to other users, deriving benefit from this activity.
Strike 3 Holdings, whose websites attract over 25 million visitors a month, is seeking compensation that could reach $359 million. The company is also pursuing a court order to prevent the distribution of its content.
In June, federal judge William Alsup ruled that the company Anthropic had the right to train models on published books without the authors’ permission.