Skip to content

Brought to you by

Dentons logo

Dentons Data

Your trusted advisor for all things digital.

open menu close menu

Dentons Data

  • Home
  • About Us

Proposed class action against Cadillac Fairview for alleged use of “facial recognition” software fails in its entirety

By Kirsten Thompson, Mark Evans, Emma Irving, and Luca Lucarini
May 20, 2025
  • Biometrics
  • Class Actions
  • Intrusion upon Seclusion
  • Litigation
  • Technology
Share on Facebook Share on Twitter Share via email Share on LinkedIn

Lawyers from Dentons Canada LLP acted for The Cadillac Fairview Corporation Limited (CFCL) in a class action proceeding advanced in BC, in relation to CFCL’s piloting of anonymous video analytics technology in 2018. In reasons for judgment issued May 15, 2025, Cleaver v the Cadillac Fairview Corporation Limited, 2025 BCSC 910 (Decision) J. Forth of the BC Supreme Court dismissed the plaintiffs’ application for certification in its entirety.

The Decision is important in several respects. First, it confirms that the finding of a privacy regulator will not, unto itself, establish the necessary basis in fact for certification – more is required. As was the case here, the absence of evidence for the existence of fundamental common issues relating to alleged invasions of privacy will not satisfy the requirements of the Class Proceedings Act, RSBC, 1996, c 50 (CPA), even on the generous evidentiary standard required at certification.

Second, the Decision sees the Court exercising a gatekeeping role with respect to the admissibility and weight that should be given to expert evidence on certification.

Finally, the Decision represents yet another instance of claims that rely on compensable harm (in this case, negligence and various breaches of Quebec law) being bound to fail for disclosing no cause of action.

Background

CFCL is a leading investor, owner and manager of commercial real estate, including shopping malls. Between May and July 2018, CFCL piloted technology (Software) in some of those malls that worked with cameras located in wayfinding directories to anonymously count the number of visitors to those malls and estimate rough gender and age ranges for each visitor (Data). The project lasted for eight weeks and ended on July 25, 2018, in response to misinformation circulating online and erroneous media reports suggesting that the Software was “facial recognition” technology.

Following the end of the pilot project, the Office of the Privacy Commissioner of Canada, the Information and Privacy Commissioner of Alberta, and the Information and Privacy Commissioner of British Columbia (Commissioners) launched a joint investigation to determine whether Cadillac Fairview was collecting and using the personal information of visitors to its malls. On October 28, 2020, the Commissioners issued the report Joint investigation of the Cadillac Fairview Corporation Limited by the Privacy Commissioner of Canada, the Information and Privacy Commissioner of Alberta, and the Information and Privacy Commissioner for British Columbia (Joint Report). In their Joint Report they found, despite evidence of the anonymous nature of the information processing, that CFCL was responsible for creating over 5 million unique biometric identifiers without notice or consent, contrary to private sector privacy legislation.

Following the Joint Report, the plaintiffs initiated civil proceedings in BC, advancing claims for intrusion upon seclusion, statutory privacy causes of action, negligence, and breach of Quebec law, and seeking to certify their claim as a class action consisting of class members in BC, Alberta, Manitoba, Ontario and Quebec. A parallel proceeding was launched in Quebec.

No personal information and no identifiable class

At the core of the claim was the question of whether CFCL had been responsible for the creation of unique, biometric identifiers – in other words, whether CFCL had collected or used any personal information.  This issue was germane not only to the alleged breaches of privacy, but also to the question of whether there was an identifiable class. CFCL argued that absent the creation of unique, identifying information by the Software, there was no possibility for putative class members to identify themselves from within the Data.

The plaintiffs and CFCL both proffered expert evidence on this issue. The Court gave the opinion of the plaintiff’s primary expert little weight, on the basis of him having failed to conduct any type of examination of the Software and, at best, offering opinions limited to the form, content and searchability of the Data.

Instead, the Court preferred the evidence of CFCL’s expert, a leading computer vision specialist, who in preparing his report after conducting experiments with the Software and ultimately determining:

  • the Software converted images of faces that it detected into a string of 128 numbers (each an Embedding Number) and generated corresponding age and gender estimations;
  • each Embedding Number was not unique to any individual and did not contain any biometric information; and
  • there was “close to zero” possibility of an individual, including a putative class member, identifying themselves from any Embedding Number, or any other part of the Data (in other words, the Data is anonymous).

In addition to their expert, the plaintiffs sought to rely on the Commissioners’ findings in the Joint Report as evidence that the Software recorded or created personal (i.e. unique) information. The Court accepted that the Joint Report was admissible but only to be used by the Court to put the facts plead into context and not for the truth of its contents (para. 66):

I accept that the OPC Report is admissible, but not for the truth of its contents. As such, I cannot rely on its findings as evidence that Cadillac Fairview collected images and biometric information without consent.

The Court thus found the plaintiffs could not rely on Joint Report’s findings as evidence that CFCL had record or created biometric or personal information.

Apparently anticipating having difficulty establishing an identifiable class, the plaintiffs amended their proposed class definition several times throughout the proceedings, ultimately settling on a proposed class that would be composed of all individuals who “viewed” a wayfinding directory within the applicable times. In light of the evidence before it, the Court found that there was no basis in fact to demonstrate that putative class members could self-identify, and further, that there was no rational relationship between the proposed class definition and the fundamental common issue, being whether a facial image of any class member was recorded and used to create biometric and personal information. On that basis, the Court held that the plaintiffs had failed to meet the identifiable class criterion set out in s. 4(1)(b) of the CPA.

Most causes of action bound to fail

While the absence of an identifiable class alone would have doomed the plaintiffs’ claim, the Court additionally found that the better part of the plaintiffs’ causes of actions were bound to fail (thus failing to satisfy the criterion under s. 4(1)(a) of the CPA); and additionally that there was no basis in fact for those that were not bound to fail (s. 4(1)(c) of the CPA). 

With respect to intrusion upon seclusion – the Court did not disturb existing Alberta jurisprudence confirming that the tort of intrusion upon seclusion does not currently exist in Alberta and thus found that claim bound to fail in Alberta. The Court also held that while the BC Court of Appeal has left open the possibility of the tort of intrusion being recognized in BC, it has not yet been, and thus held that the plaintiff’s claim for intrusion upon seclusion was also bound to fail in BC.

With respect to negligence– the Court agreed with CFCL that the law is clear that claims for mental injury that are limited to upset, disgust, anxiety, agitation or other mental states falling short of injury are not compensable damages in a negligence claim; and found negligence also bound to fail.

With respect to the statutory breach torts (BC and Manitoba) – the Court found that the claims based on CFCL’s alleged violation of putative class members privacy by using their likenesses for the purposes of advertising or promotion were bound to fail, with the plaintiffs’ allegations in this regard amounting to little more than bald assertions unsupported by any material facts.

With respect to the breaches of Quebec law – as with negligence, the Court found the law in Quebec clear that fault alone does not cause damage and, absent compensable damage, a cause of action under the Civil Code of Quebec bound to fail.

No basis in fact for the core common issues

The plaintiffs sough to certify numerous common issues, chief among them being common issues about whether CFCL had recorded facial images of mallgoers, whether it had used those facial images to create biometric and personal information, what uses had been made of that information, and what additional uses could have been made of that information.

The Court found that the evidence proffered by the plaintiff’s expert on these points to be essentially a reiteration of the findings in the Joint Report and concluded that “[s]ince the OPC Report is not admissible for the truth of its contents, I would be in error if I indirectly accepted the truth of its contents by relying on an expert opinion that is based on them”.

With the balance of the plaintiff’s expert’ opinion being limited to inadmissible speculation, the Court thus determined that the only evidence before it on these points was that of CFCL’s expert who, as noted above, found that the Software had unequivocally not created biometric or personal information.

Accordingly, the Court held there to be no basis in fact for the existence of these core common issues, and correspondingly, no basis in fact for any of the other common issues, all of which were contingent on CFCL’s creation of biometric or personal information. The plaintiffs thus failed to satisfy the criterion at s. 4(1)(c) of the CPA.

Not a preferable procedure

Finally, the Court found a class proceeding would not have been the preferable procedure (s. 4(1)(d) of the CPA), because:

  • there was a near total absence of common issues;
  • neither of the proposed representative plaintiffs had submitted any evidence of actual or compensable harm; and
  • there was no need for behaviour modification (CFCL has ceased the pilot project and had no ongoing access to the Data).

On the basis of the foregoing, the Court found that to deploy scarce judicial resources on the plaintiff’s claim, would be the “antithesis of judicial economy and would not provide meaningful access to justice.”

Dentons was pleased to represent The Cadillac Fairview Corporation Limited in this matter.


For more information on this topic, please contact Kirsten Thompson, Mark Evans, Emma Irving, Luca Lucarini or other members of the Dentons Privacy and Cybersecurity group.

Share on Facebook Share on Twitter Share via email Share on LinkedIn
Subscribe and stay updated
Receive our latest blog posts by email.
Stay in Touch
Class actions, Litigation
Kirsten Thompson

About Kirsten Thompson

Kirsten Thompson is a partner and the national lead of Dentons’ Privacy and Cybersecurity group. She has both an advisory and advocacy practice, and provides privacy, data security and data management advice to clients in a wide variety of industries.

All posts Full bio

Mark Evans

About Mark Evans

Mark Evans is the National Practice Group Leader for our Litigation and Dispute Resolution group in Canada. He is also a member of the Canada Firm’s National Management Committee, the Global co-leader for the Firm’s Fraud and Asset Recovery Group and a member of the global litigation leadership team.

Full bio

Emma Irving

About Emma Irving

Emma Irving is co-leader of the national Class action group and partner in Dentons’ Litigation and Dispute Resolution group in Vancouver. She represents national and regional clients in a variety of different commercial and regulatory disputes. Emma’s practice encompasses most areas of commercial litigation with a focus on contract and transactional disputes, administrative law, personal injury defence, class actions and product liability.

All posts Full bio

Luca Lucarini

About Luca Lucarini

Luca Lucarini is an associate in our Litigation & Dispute Resolution and Transformative Technologies and Data Strategy groups. Luca acts for clients on a variety of regulatory, commercial and civil litigation matters, with a particular emphasis on privacy and health law.

All posts Full bio

RELATED POSTS

  • Artificial Intelligence
  • Litigation
  • Technology

Airline ordered to compensate a B.C. man because its chatbot provided inaccurate information

By Kirsten Thompson
  • Litigation
  • Privacy

Breach-related privacy claims against company struck where intentionality absent

By Kirsten Thompson
  • Class Actions
  • Privacy

Certification of Breach Class Action Denied in Absence of Provable Losses, Commonality

By Kirsten Thompson and Mike Schafler

About Dentons

Redefining possibilities. Together, everywhere. For more information visit dentons.com

Grow, Protect, Operate, Finance. Dentons, the law firm of the future is here. Copyright 2023 Dentons. Dentons is a global legal practice providing client services worldwide through its member firms and affiliates. Please see dentons.com for Legal notices.

Categories

Subscribe and stay updated

Receive our latest blog posts by email.

Stay in Touch

Dentons logo in black and white

© 2025 Dentons

  • Legal notices
  • Privacy policy
  • Terms of use
  • Cookies on this site