Elsevier has joined a class-action lawsuit against Meta, accusing the tech giant of reproducing copyrighted works to develop its Llama AI model. The publisher is the first science publisher to take legal action over scraped research papers in this context, making it a pivotal moment for how AI companies source training data.
What the lawsuit alleges
The suit, filed on May 11, claims Meta used copyrighted materials from Elsevier's journals without permission while building the Llama model. It argues that reproducing those papers in training datasets amounts to copyright infringement. The case is part of a broader class action, but Elsevier stands alone among academic publishers in joining so far.
📊 Market Data Snapshot
This isn't just a publishing fight. The lawsuit targets the fundamental 'scrape-and-train' model that many tech projects rely on — including blockchain-based AI protocols that pull data from the open web. If the court finds that using scraped research papers violates copyright, it could force entire categories of AI development to rethink how they acquire data. Crypto projects that depend on freely ingested information could face similar legal exposure.
What happens now
No court date has been set yet. The case adds to growing regulatory pressure on AI data practices, especially as the EU AI Act's research exemptions begin clashing with copyright enforcement. For now, all eyes are on whether Meta will argue fair use or seek a settlement. A ruling could take months, but the precedent is already being watched by both traditional AI firms and decentralized networks.
