On October 6, a group of major Canadian media outlets filed a lawsuit against Anthropic in an Ontario court. The plaintiffs, including The Globe and Mail and Postmedia, claim the AI firm used their journalistic work to train the Claude model without permission.
Postmedia and The Globe and Mail target Claude's training data
A coalition coonsisting of Postmedia, Torstar, The Globe and Mail, and The Canadian Press has alleged that Anthropic engaged in copyright infringement to build its large language model, Claude. According to The Canadian Press, the plaintiffs argue that Anthropic used their proprietary content to develop its AI services without securing the necessary licenses.
The statement of claim asserts that Anthropic has been "substantially and unjustly enriched" by capitalizing on the commercial success of its models and raising significant capital. The news organizations argue that this growth came at their direct expense, as Anthropic allegedly leveraged their journalistic output to create a profitable suite of AI tools without providing compensation.
The allegation of circumventing technological protection measures
Beyond simple copyright infringement, the lawsuit claims that Anthropic deliberately bypassed technical barriers. As reported by The Canadian Press, the news outlets allege that the AI company circumvented technological protection measures specifically designed to prevent the unauthorized copying of their content.
The claim further suggests that Anthropic breached the terms of service of the various news websites. The publishers allege that Anthropic continues to scrape enormous amounts of text data from the internet and their specific sites on an ongoing basis, even when responding to individual user prompts, to keep the Claude model updated.
A repeat of the 2024 OpenAI legal battle
This legal action is not an isolated incident but part of a broader strategy by Canadian publishers to protect their intellectual property. The same group of companies, joined by CBC/Radio-Canada, launched a similar lawsuit against OpenAI in 2024. That case, which also centers on the unauthorized use of news content by AI systems, remains unsettled.
This pattern suggests that Canada's largest media entities are moving toward a systemic legal challenge against the "scrape-first, ask-later" mentality of AI developers. By targeting both OpenAI and Anthropic, the publishers are attempting to establish a legal precedent that mandates licensing agreements for the ingestion of journalistic data.
The failure of Canada's national AI strategy on copyright
The move to the Ontario court reflects a growing frustration with federal inaction. news publishers have repeatedly urged the Canadian government to clarify how AI systems should handle copyrighted material, but the government's national AI strategy failed to provide a concrete framework for this issue.
Because the regulatory environment remains ambiguous, pbulishers feel compelled to seek remedies through the judiciary. They argue that without government intervention or mandatory licensing, AI firms will continue to erode the economic viability of professional journalism by using the very content they are potentially replacing.
The undisclosed damages and Anthropic's silence
Despite the gravity of the claims , several key details remain missing from the public record. The statement of claim does not disclose the exact amount of damages the publishers are seeking from Anthropic in the Ontario court.
Furthermore, there is currently only one side to the story. Anthropic has not issued a public statement or responded to requests for comment regarding the allegations of scraping and the breach of terms of service. it remains to be seen whether Anthropic will argue "fair dealing" or attempt to negotiate a licensing deal similar to those seen in other jurisdictions.
Comments 0