Meta Platforms, the parent company of Facebook, paused its efforts to license books for training its AI models in early April 2023 due to logistical and practical challenges. This decision has come under scrutiny as part of the ongoing copyright infringement lawsuit Kadrey v. Meta Platforms, which accuses the company of using copyrighted and pirated materials without proper authorization to train its AI systems, including the LLaMA model series.
Table of Contents
Reasons for Pausing Licensing Efforts
Meta’s attempts to negotiate licensing agreements with publishers faced significant hurdles:
Lack of Engagement
Meta’s outreach to publishers was met with slow responses and limited interest. According to Sy Choudhury, who leads Meta’s AI partnerships, many cold-call attempts to establish contact with publishers were unsuccessful.
Rights Issues
Many publishers, particularly in the fiction category, did not hold the necessary rights to license their books for AI training. This complicated negotiations, as Meta would have had to engage directly with individual authors, a time-consuming process.
Logistical Setbacks
The company cited timing and logistical challenges as additional reasons for halting discussions.
Allegations in the Lawsuit
The lawsuit, filed by authors including Richard Kadrey, Sarah Silverman, and Ta-Nehisi Coates, alleges that Meta used unauthorized sources for training its AI models:
Use of Pirated Books
Plaintiffs claim that Meta accessed “shadow libraries” containing pirated e-books and cross-referenced them with copyrighted works to evaluate licensing feasibility.
Copyright Infringement
The plaintiffs argue that this practice violates copyright laws and undermines the intellectual property rights of authors.
Broader Implications
This case is part of a larger wave of lawsuits against AI companies like OpenAI and Anthropic over the use of copyrighted materials for training generative AI systems. While tech companies often argue that such use falls under “fair use,” copyright holders contend that it constitutes unlawful copying and threatens their livelihoods.
Meta’s decision to pause book licensing efforts reflects broader challenges in balancing AI development with intellectual property rights. As litigation continues, the outcome could have significant implications for how AI companies source training data in the future.