Executive Summary: Key Legal and Evidentiary Issues

• Certification was denied due to lack of an identifiable class and absence of common issues.

• Plaintiffs failed to show that biometric or personal data was captured or stored by the software.

• Expert evidence relied heavily on a non-admissible privacy commissioner report, undermining its probative value.

• The court held that speculative harm and emotional distress were not sufficient to ground tort or negligence claims.

• Intrusion upon seclusion was only potentially viable in Ontario and Manitoba, not in British Columbia or Alberta.

• Use of facial embeddings was not proven to constitute identifiable or biometric personal information under law.

Facts and outcome of the case

Background and factual allegations

In this case, the plaintiffs, Joshua Cleaver and Curtis Kieres, alleged that Cadillac Fairview unlawfully captured and processed biometric data from mall visitors using hidden cameras embedded in digital wayfinding directories. These devices were equipped with software provided by a third party, MappedIn Inc., and were deployed in multiple shopping malls across Canada during a pilot project in 2018.

The plaintiffs claimed that facial images of mall visitors were captured without consent, transformed into numerical data (so-called "embeddings"), and retained for analysis. The plaintiffs alleged this amounted to violations of their privacy rights under various provincial and federal statutes, and they sought certification of a national class action covering multiple jurisdictions, including Quebec.

Legal issues and analysis

Madam Justice Forth analyzed the plaintiffs’ application under the Class Proceedings Act, focusing on whether the criteria for certification under section 4(1) were satisfied. The plaintiffs put forward several causes of action, including intrusion upon seclusion, negligence, and breaches of privacy legislation such as the BC Privacy Act, Manitoba Privacy Act, Alberta PIPA, and PIPEDA. They also alleged violations of Quebec’s Civil Code and privacy statutes.

The key evidentiary issue was whether facial images were recorded and whether the resulting data could be considered biometric or personal information. The plaintiffs’ expert, Jason Frankovitz, relied heavily on the findings of the federal and provincial privacy commissioners, which the court ruled inadmissible for truth of content. Additionally, his report was criticized for being speculative and lacking independent technical analysis.

Conversely, the defendant's expert, Professor Richard Zhang, conducted experiments using the actual data and concluded that the embeddings were not identifiable, not biometric in nature, and could not be reversed to reveal original facial images. His findings played a decisive role in undermining the plaintiffs’ assertions.

Outcome of the decision

The court dismissed the plaintiffs' application for class action certification. Justice Forth found that the plaintiffs failed to demonstrate:

• That the facial images were recorded in the way alleged;

• That the resulting data constituted personal or biometric information;

• That there was an identifiable class capable of self-identification;

• That the common issues were provable on a class-wide basis;

• That a class action was the preferable procedure.

The court held that while some statutory and tort claims may be arguable in theory, they were not supported by sufficient facts to proceed. The speculative and non-compensable nature of the alleged harms further weakened the plaintiffs’ position. Although some causes of action were not struck outright, the failure to satisfy the core requirements under the Class Proceedings Act meant that the proceeding could not advance.

Costs and final disposition

No costs were awarded at the time of judgment. The court invited the parties to submit written submissions on costs within 60 days or to request an oral hearing within 30 days. No damages were awarded, as the case was dismissed at the certification stage without reaching the merits.