Five major publishers and bestselling novelist Scott Turow have filed a class-action lawsuit against Meta Platforms Inc. and its founder Mark Zuckerberg, alleging the company used millions of copyrighted books and journal articles to train its Llama artificial intelligence model without authorization or compensation.
The complaint, filed in federal court, accuses Meta of using works obtained through pirated sites including LibGen and Sci-Hub. The filing claims Zuckerberg "personally authorized and actively encouraged the infringement" in building the AI system. Meta has not publicly responded to requests for comment on the lawsuit.
The case is part of a broader wave of litigation against AI companies over training data. Last month, Anthropic agreed to a $1.5 billion settlement with writers over similar copyright claims. Other major technology firms including OpenAI and Google face related lawsuits from publishers and authors.
What the Left Is Saying
Democratic lawmakers and progressive advocacy groups have largely sided with the publishers and authors in these cases. Senator Amy Klobuchar of Minnesota, who has led efforts on AI transparency legislation, said the Meta lawsuit highlights "the need for clear rules about how AI companies can use creative work."
"Writers spend years developing their craft, and they deserve to know if their work is being used to build systems that could eventually replace them," Klobuchar stated at a recent Senate Commerce Committee hearing. The has proposed legislation requiring AI companies to disclose training data sources.
The Authors Guild, which represents thousands of professional writers, filed an amicus brief supporting the lawsuit. CEO Mary Rasenberger said the case goes to "the foundation of creative industries" and called for federal legislation establishing licensing requirements for AI training data.
Progressive economic policy advocates argue that without intervention, AI companies will continue consolidation of creative works into proprietary systems, concentrating power among a handful of technology firms while displacing millions of writers, journalists, and researchers.
What the Right Is Saying
Technology industry groups and conservative commentators have defended AI training practices as falling under fair use doctrine. The Computer & Communications Industry Association, which represents Meta and other tech companies, argues that AI training on publicly available works is analogous to how humans learn from reading.
"Every student who reads books to learn writing is doing what these AI systems do," said CCIA President Matt Schruers in a statement. "The law has long recognized that transformative use does not require permission or payment."
Republican lawmakers have been more hesitant to support restrictive AI regulations, citing concerns about U.S. competitiveness against China. Senator John Thune of South Dakota, who chairs the Senate Commerce Committee, has advocated for voluntary industry standards rather than federal mandates.
"We need to be careful not to regulate this technology out of existence before American companies can fully develop it," Thune said during floor remarks last month. Several Republican members have introduced legislation that would limit liability for AI companies using copyrighted material in training, provided they take basic steps to address infringement claims.
Conservative legal scholars argue that expansive copyright protections for AI training could stifle innovation and create barriers to entry that favor established publishing houses over smaller technology developers.
What the Numbers Show
The publishing industry generated approximately $68 billion in revenue in 2024 according to the Association of American Publishers, with print books accounting for roughly $23 billion. The sector employs more than 500,000 workers directly, with an additional estimated 1.2 million jobs in related industries like printing, distribution, and retail.
AI-generated book titles on Amazon increased by over 900% between 2023 and 2025 according to data from the platform, though company officials note that many are low-volume self-publications rather than commercially significant works.
The lawsuit specifically cites research showing that Llama was able to produce "convincing renditions" of individual authors' writing styles when prompted, raising questions about derivative work protections. Industry analysts estimate AI-generated content could eventually represent 10-20% of new book releases if current legal frameworks remain unchanged.
Congressional Budget Office estimates suggest comprehensive AI copyright legislation could add $2-4 billion annually in licensing revenue to the publishing sector over a decade, though such projections carry significant uncertainty given rapidly evolving technology and market conditions.
The Bottom Line
The Meta lawsuit represents one of the most significant intellectual property cases in the artificial intelligence era. A ruling against Meta could establish precedent requiring AI companies to obtain licenses or explicit permissions before using copyrighted works for training, fundamentally reshaping how AI systems are developed.
Court proceedings are expected to take 18-24 months before any substantive rulings on fair use doctrine apply to AI training practices. In the meantime, Congress faces increasing pressure from both creative industry stakeholders seeking stronger protections and technology companies warning against overregulation.
What happens next: The judge must first decide whether to certify the class action covering all authors whose works were allegedly used. If certified, settlement negotiations or trial would follow. Separately, bipartisan Senate legislation addressing AI copyright is expected to receive committee votes by spring.
Key things to watch: Meta's formal response to the lawsuit; any potential settlement discussions before trial; and how the Thune-Klobuchar compromise negotiations on AI legislation progress in coming months.