Canadian Authors Sue Big Tech Giants for Alleged Copyright Infringement in AI Training

Canadian author J.B. MacKinnon has launched a class-action lawsuit against Nvidia, alleging the tech giant used his and other Canadian writers' works to train AI without paying licensing fees or securing consent. MacKinnon claims the companies took steps to conceal copyright infringement and trained LLMs to respond in a misleading way. The case highlights the tension between human writers and AI-generated content.

Title: Canadian Author J.B. MacKinnon Files Class-Action Lawsuit Against Nvidia

Canadian author J.B. MacKinnon has launched a class-action lawsuit against Nvidia Corp., alleging that the tech giant used his and other Canadian writers' works to train AI models without paying licensing fees or securing consent. The lawsuit, filed in British Columbia Supreme Court, seeks to represent all holders of Canadian copyrights whose works were used by the company to train its large language models.

The lawsuit, which is one of several similar actions filed against tech giants like Meta Platforms and Anthropic, accuses Nvidia of relying on a pirated dataset containing copyrighted material, including works by MacKinnon. The company is alleged to have used software to remove copyright management information from the books, making it difficult to trace the original authors.

Nvidia's use of the dataset, known as the Pile, which contains a subset called Books3, is at the center of the allegations. The dataset was taken from Bibliotik, a mix of fiction and nonfiction books, including works by Canadian authors with active copyright protection. The lawsuit claims that Nvidia downloaded the dataset without paying a licensing fee and used it to train its large language models (LLMs), which has helped grow the company into the world's largest by market capitalization.

The lawsuit seeks several remedies, including an injunction preventing Nvidia from infringing on copyrights, an accounting of profits, restitution to the owners, and punitive damages for the "repeated, willful and knowing infringement" of copyrights. Co-counsel Reidar Mogerman emphasized the importance of these cases for protecting creators' rights at a time when tech companies are arguing against paying for content due to AI's transformative nature.

The case highlights the ongoing tension between human writers and AI-generated content, as well as the legal challenges surrounding AI training and the use of copyrighted materials. While the Canadian courts are seen as more protective of copyright than in the U.S., the success of these cases remains uncertain.

References
[1] https://vancouversun.com/news/bc-author-canadian-class-action-lawsuit-facebook-nvidia-ai
[2] https://www.plagiarismtoday.com/2025/07/30/3-strange-ways-ai-is-improving-plagiarism/