Meta’s CEO, Mark Zuckerberg, allegedly allowed developers to use pirated content for training Meta’s AI models. Developers reportedly removed copyright notices from the material. Lawyers for several prominent U.S. authors accuse Meta of this in a lawsuit in a California court.
Recently released documents cite statements from Meta employees and internal correspondence. According to these documents, Meta’s AI team received approval to use data from LibGen for training the Llama models after escalating the request to “MZ,” which stands for Mark Zuckerberg. The lawyers describe LibGen as a collection of pirated copyrighted works. All decision-makers, including Zuckerberg, were aware that the data was pirated.
The Meta developers were initially hesitant to use the data. One employee noted, “Using file-sharing networks on a company laptop doesn’t feel right.” Meta reportedly admitted to removing copyright notices from e-books and scientific articles sourced from LibGen. The authors’ lawyers argue that the company wanted to prevent AI-generated responses from including references to copyrighted material. According to the documents, Meta developers had to upload copyrighted material to file-sharing networks to download other content.
In the lawsuit, U.S. authors Sarah Silverman, Richard Kadrey, and Christopher Golden accuse Meta of illegally using their books for AI model training. They also claim that the AI models’ responses violate copyright laws. In September 2023, the court dismissed most of the authors’ claims, except for the claim that training the AI models with copyrighted works infringed copyright. Meta has not responded to a request for comment.
This situation highlights the ongoing tension between technology companies and copyright laws. As AI technology advances, the use of copyrighted material for training purposes becomes a contentious issue. The debate centers around whether using such material without permission constitutes fair use or copyright infringement.
Authors and creators argue that their work is being used without compensation, potentially affecting their income and intellectual property rights. On the other hand, tech companies argue that AI training requires vast amounts of data, and accessing such data can be challenging without using existing materials.
The case against Meta is part of a broader trend where creators are increasingly taking legal action against tech companies. They seek to protect their rights and ensure they are fairly compensated for the use of their works. This legal battle could set a precedent for how AI training data is sourced and used in the future.
Despite the court dismissing most claims, the case continues to draw attention to the ethical and legal implications of AI development. It raises questions about the responsibilities of tech companies to respect intellectual property rights while pushing the boundaries of innovation.
As the case unfolds, it may influence how other tech companies approach AI training. It could lead to more stringent guidelines and practices for sourcing training data. The outcome may also impact how copyright laws are interpreted in the context of AI and digital content.
While the legal process is ongoing, the case serves as a reminder of the importance of balancing technological advancement with respect for creators’ rights. It highlights the need for clear regulations and ethical standards in the rapidly evolving field of artificial intelligence.