Meta Secures Victory in Copyright Lawsuit Over AI Training Data

 

In a significant ruling on June 25, 2025, Meta, the parent company of Facebook, won a federal lawsuit in the U.S. brought by 13 prominent authors, including Sarah Silverman and Ta-Nehisi Coates. The authors claimed Meta illegally used their copyrighted works to train its AI model, Llama. This marks the second favorable decision for AI companies in similar copyright disputes within a week, spotlighting the ongoing clash between technological innovation and intellectual property rights.

Judge Vince Chhabria dismissed the case, stating that the authors presented “flawed arguments.” However, he emphasized that the ruling does not broadly endorse Meta’s use of copyrighted materials for AI training. Instead, it highlights the specific shortcomings of this case, leaving room for future legal challenges with stronger arguments. Meta welcomed the decision, arguing that fair use of copyrighted materials is critical for driving transformative AI innovations that enhance creativity and productivity.

The authors’ legal team contended that Meta sourced their works from pirated online repositories, labeling it a “massive copyright infringement.” They argued that Meta should have licensed or paid for the content, given the substantial profits AI technologies generate. In his 40-page ruling, Chhabria dismissed claims that adhering to copyright laws would hinder technological progress, noting that companies like Meta have the means to fairly compensate creators for their work.

This ruling underscores the growing tension between AI development and the rights of content creators. While Meta celebrates this win, the judge’s remarks suggest that similar cases could resurface with refined arguments. As AI continues to reshape industries, the legal system will play a pivotal role in defining how companies balance innovation with respect for intellectual property, keeping this issue at the forefront of global discussions.