Zuckerberg was reportedly aware that Meta trained its AI model on pirated works

The core revelation: Meta CEO Mark Zuckerberg approved the use of Library Genesis (LibGen), a known pirated content repository, to train the company’s Llama 3 AI model, according to newly unsealed court documents.

Key details of the disclosure: Internal communications revealed through a class-action lawsuit show Meta executives discussing the company’s deliberate use of unauthorized copyrighted material.

• Sony Theakanath, Meta’s director of product management, confirmed in an email that Zuckerberg approved LibGen’s use for AI training
• The company explicitly planned to keep its use of LibGen confidential
• Meta employees discussed methods to remove copyright indicators from the pirated content
• Internal discussions revealed concerns about downloading pirated content from corporate devices

Legal context: A class-action lawsuit filed by authors Christopher Golden, Richard Kadrey, and comedian Sarah Silverman alleges unauthorized use of their copyrighted work.

• The documents were unsealed by Judge Vince Chhabria of the U.S. District Court for Northern California
• Meta’s legal team had previously argued that their use of text for AI training fell under fair use provisions
• Zuckerberg reportedly acknowledged in a deposition that such piracy would raise “lots of red flags”

Corporate strategy and risk assessment: Meta executives weighed the benefits against potential backlash while implementing this controversial decision.

• Internal communications cited performance benchmarks as justification for using LibGen
• Documents referenced rumors that competitors like OpenAI and Mistral AI were also using the library
• Executives acknowledged potential legislative risks, particularly in the US and EU
• The company developed specific “mitigations” to address potential fallout

Industry implications: This revelation comes at a critical time for AI development and copyright law.

• Meta announced a 5% workforce reduction targeting “lowest performers” (approximately 3,600 workers)
• The case could set important precedents for numerous other AI-related copyright lawsuits
• The controversy highlights the tension between rapid AI development and intellectual property rights

Analyzing the deeper impact: This controversy exposes a fundamental contradiction in the AI industry’s approach to training data – while companies need vast amounts of high-quality content to develop effective AI models, their methods of obtaining this content often conflict with established intellectual property rights, potentially setting up a long-term conflict between content creators and AI developers.