The $5 Billion Question: Did Facebook’s CEO Know?
A bombshell lawsuit filed by publishers and author Scott Turow alleges that Mark Zuckerberg personally authorized and encouraged Meta to scrape billions of copyrighted books, articles, and images without permission to train its AI systems. This isn’t just another copyright dispute—it’s a direct challenge to the legitimacy of generative AI training on public content, and it implicates one of the world’s most powerful tech leaders in a deliberate act of intellectual property theft. The plaintiffs claim internal Meta documents show executives greenlit a project codenamed 'Seventeen' that systematically harvested vast troves of unlicensed material from the internet, including entire libraries of academic journals, news archives, and literary works.
How ‘Seventeen’ Became Meta’s Secret Weapon
According to discovery materials cited in the complaint, 'Seventeen' was launched in early 2023 as a high-priority initiative to build Meta’s large language model capabilities ahead of competitors like OpenAI and Google. Rather than relying on licensed datasets or user-generated content, the project extracted terabytes of text and visual data from websites such as Project Gutenberg, Reddit, and thousands of publisher-hosted digital archives. Internal emails reveal concerns about legality were dismissed with statements like, 'We’re not doing anything wrong here,' and 'Zuck wants this prioritized.' The scale was staggering: over 10 billion pages indexed, including paywalled content and subscription-based sources.
The Legal Firestorm Ignites
Publishers including Penguin Random House, Wiley, and the Authors Guild argue that Meta’s actions constitute willful infringement under U.S. copyright law, especially given the absence of opt-out mechanisms for web crawlers. They point to Meta’s own privacy policy, which once promised not to use personal information 'to train our AI models,' yet now claims broad rights to all publicly accessible data. Scott Turow, a former president of the Mystery Writers of America, is particularly incensed—his novel *Limitations* was scraped en masse, despite his public stance against AI using his work without consent. His lawsuit joins dozens of others across the publishing industry, collectively seeking hundreds of millions in damages.
Why This Changes Everything
If true, Meta’s strategy exposes a systemic flaw in how major AI companies justify training data sourcing. Generative AI models are only as good as their data, and if that data is built on stolen creativity, then the very output of those models—from news summaries to fictional stories—becomes legally tainted. More importantly, the lawsuit threatens to set a precedent: if Zuckerberg knew and approved mass scraping, could any AI firm plausibly claim ignorance? It also raises urgent questions about corporate accountability in the AI race, where speed often trumps ethics. As publishers push back, the outcome may determine whether AI development operates under guardrails or in a legal gray zone.