Anne Merminod
Torys, Montreal
amerminod@torys.com

Introduction

Canada’s litigation landscape is rapidly evolving as artificial intelligence (AI) technologies become integral to commercial operations. Over the past year, the country has witnessed a surge in class actions targeting AI developers and users, especially in the realms of copyright infringement and data privacy. These developments are particularly significant for international dispute resolution, given Canada’s well-established class action regime and its attractiveness to plaintiffs seeking redress against global technology companies.

Canada’s class action expertise: Quebec and BC as hot spots

Canada is deeply familiar with class action proceedings, having developed a sophisticated and mature regime over several decades. Each year, Canadian courts see around 200 new class action filings,^[1] with the majority concentrated in Quebec, British Columbia (BC) and Ontario. In 2024, Quebec led the way with 67 filings (about 35 per cent of the national total), closely followed by BC with 63 filings (32.5 per cent), and Ontario with 48 filings (24.7 per cent).

Both Quebec and BC have become recognised hotspots for class actions, thanks to their plaintiff-friendly certification processes and no-cost regime. Quebec’s courts routinely authorise class actions that meet a relatively low threshold, while BC’s streamlined certification process and the absence of a predominance test make it a preferred venue to Ontario. Ontario’s popularity has decreased following the 2020 amendments to its Class Proceeding Act.

Based on the pace of filings since January, we expect 2025 to be another record-setting year, with even more class actions initiated – especially in Quebec and BC – which continue to attract the highest volume of new cases.

Generative AI and copyright: a new litigation frontier

The rise of generative AI – systems that create text, images, and other content by training on vast datasets – has given birth to a new category of litigation. Plaintiffs are increasingly alleging that AI models are trained on copyrighted works without authorisation, raising complex questions about fair use, licensing and creator rights.

In Quebec, an author has launched four proposed class actions against major tech companies, alleging unauthorised use of Canadian authors’ books to train AI models. In BC, an author has also launched four proposed class actions against the same tech companies. A coalition of major Canadian news outlets also filed suit against OpenAI, claiming their content was scraped and copied to train AI systems. Very recently, another author alleging unauthorised use of Canadian authors’ books to train AI models targeted two other tech giants with similar allegations. Federal court cases alleging similar issues have been launched as well. These recent filings shows that the trend in AI class actions is following the United States surge.

Expansion beyond copyright: data privacy and insurance

AI class actions in Canada increasingly intersect with data privacy. Parallel cases have also been filed in federal court, raising similar allegations.

The wave of AI and data-driven litigation is further illustrated by the proposed class actions, which have been initiated in both Quebec and the Federal Court around collecting and processing biometric information – specifically, images of Canadians – without consent, in violation of privacy laws.

Insurers deploying AI for claims processing, underwriting, and customer service also face litigation risks around bias, misrepresentation, breach of contract and privacy breaches.

Regulatory risk is also evolving, with provincial laws and guidance from regulators such as Office of the Superintendent of Financial Institutions (OSFI) and Quebec’s Autorité des marchés financiers (AMF) shaping AI governance.

Legal factors fuelling the rise of class actions

Several features of Canadian law make the jurisdiction particularly conducive to AI-related class actions:

• Low certification thresholds: Canadian courts maintain a relatively low bar for certifying class actions. In Quebec and the Federal Court, authorisation/certification rates approached 80 per cent in 2024.
• Statutory damages: The Canadian Copyright Act allows plaintiffs to elect statutory damages – ranging from CAD$500 to CAD$20,000 per infringed work for commercial use – without the need to prove actual losses. The privacy regulations have similar statutory damages.

Conclusion

Canada’s deep familiarity with class actions has positioned it as a hotbed for AI-related litigation. For international litigators and commercial parties, understanding the Canadian approach is essential for effective risk management and dispute resolution.

Anne Merminod will be speaking on the panel ‘Class actions in the era of artificial intelligence’ at the IBA Annual Conference in Toronto on 5 November 2025.