The ongoing debate over large language models (LLMs) and their use of open-source information has reached a new stage. Canadian publishers are now suing OpenAI, alleging copyright infringement. Media outlets have long been concerned about the unauthorized use of their materials, and many have sought to address these issues in court.
As Bloomberg reports, the plaintiffs include Torstar Corp., Postmedia Network Canada Corp., Globe and Mail Inc., Canadian Press, and CBC/Radio-Canada. These companies filed a lawsuit in the Ontario Superior Court, seeking unspecified damages for OpenAI’s alleged misuse of their text materials during the training of its language models. The claim states: “OpenAI earns and benefits from the use of this content without receiving official permission or compensating the copyright holders.” According to the plaintiffs, their publications represent a significant portion of journalism content in Canada.
OpenAI’s Response to Allegations
In response, OpenAI asserts that its models are trained on “publicly available data, in a fair use manner, and in accordance with internationally recognized copyright principles.” The company also claims to collaborate with publishers to develop acceptable methods for processing and displaying content. OpenAI emphasizes efforts to credit original sources and states that materials can be removed at the request of copyright holders.
Copyright Battles Continue
This lawsuit follows a similar case from late 2023, when The New York Times sued OpenAI, accusing the company—and its close partner Microsoft—of using millions of copyrighted articles to train AI models, adds NIX Solutions. News Media Canada CEO Paul Deegan accused OpenAI of “ripping off journalists while illegally and unjustifiably substantially enriching itself.”
As the legal landscape evolves, tensions between AI developers and content creators remain high. We’ll keep you updated on how these cases unfold and whether they lead to significant changes in copyright practices for LLMs.