The publishers argue that OpenAI utilizes their news content to train ChatGPT, which constitutes a violation of copyright and allows the company to profit from the material without permission or compensation.
OpenAI contested the jurisdiction of the court by raising a plea to the jurisdiction and arguing the matter should instead be tried before an American tribunal. The Ontario Court of Appeal subsequently rejected the appeal.
The court’s decision specifies that OpenAI presented no evidence demonstrating that a judgment rendered in Canada would conflict with existing United States laws, which would complicate its execution.
This lawsuit, which has been ongoing, represents the first Canadian case addressing the practice of using copyright-protected content to train generative artificial intelligence systems. Earlier this year, Postmedia, Torstar, The Globe and Mail, and La Presse Canadienne initiated a separate action against artificial intelligence firm Anthropic.
Information and cultural sector associations have called on the federal government to address how AI systems use copyright-protected content, though Ottawa has indicated no intention to act or to update the copyright law.
Parallel legal challenges also involve the entertainment industry. Their core point is simple: media businesses secure licenses to use external material routinely, and artificial intelligence developers should not be exempt from this obligation merely because compliance poses logistical challenges.
An amicus brief allows an outside party to offer arguments to the court. The submitted document highlights a sector that regularly manages permissions for literature, musical compositions, acted performances, and visual media, frequently coordinating multiple individual and corporate stakeholders within a single venture.
Creative Artists Agency and other creative-industry organizations are also supporting the filing. The complaint at the center of the case concerns copyrighted journalism used to train AI systems. While the involvement of actors widens the debate to include industry practitioners, this backing does not imply any judicial determination that the likeness, vocal identity, or work of every individual signatory was actually misappropriated here.
In a publication dated October 30, the firm highlighted perspectives from external organizations backing that stance, characterizing fair use as a mechanism that balances public benefit with technological progress. This establishes the central legal question: do the disputed applications fall under permitted copyright exemptions without authorization, or do the challenges raised by copyright owners mandate a different ruling?
Because artificial intelligence models are often trained prior to public release, creators are frequently forced to dispute consent after their content has already been harvested. Asking first would give them a chance to refuse or negotiate terms before that use happens.
Human Consent aims to simplify the expression of these choices via a public directory and a machine-readable protocol, enabling individuals to specify if usage is allowed, restricted by conditions, or banned entirely.
Granting consent for a specific movie does not clarify whether an individual endorses artificial intelligence training, a digital replica of their performance, or an entirely separate future production, given that separate agreements govern such terms. Blanchett’s group wants the permission conversation to happen before the work is taken, while OpenAI argues that the law already allows the training in dispute.