Cadillac Fairview Collected Five Million Shopper Images Using Undisclosed Facial Recognition in Canadian Malls
Over five million facial representations were captured and analyzed without knowledge or consent from shoppers at 12 malls across five provinces — one of the largest documented undisclosed biometric data collection operations in Canada.
Narrative
Cadillac Fairview, one of Canada’s largest commercial property owners, embedded small cameras inside digital directory kiosks at 12 shopping malls across five provinces — Ontario, Quebec, Alberta, British Columbia, and Manitoba. These included high-traffic properties such as Toronto’s Eaton Centre and Montreal’s CF Carrefour Laval. The cameras captured images of shoppers without their knowledge or consent — over five million numerical facial representations were generated — and facial recognition software analyzed the images to estimate each person’s age and gender.
A joint investigation by the federal Privacy Commissioner, Alberta’s Information and Privacy Commissioner, and British Columbia’s Information and Privacy Commissioner — with Quebec’s Commission d’accès à l’information collaborating separately — found that Cadillac Fairview violated the Personal Information Protection and Electronic Documents Act (PIPEDA). The investigation revealed that a third-party technology provider had retained approximately five million numerical facial representations on a decommissioned server. The commissioners found the company had failed to obtain meaningful consent for the collection of sensitive biometric information and had not been transparent about the facial recognition capabilities embedded in the kiosks.
The investigation concluded that shoppers had no reasonable expectation that visiting a mall would result in the capture and analysis of their biometric data. The cameras were inconspicuous and no signage informed visitors of the facial recognition technology. The commissioners recommended that Cadillac Fairview delete the collected data and obtain express consent before any future use of such technology. Cadillac Fairview subsequently confirmed it had deleted the data and removed the cameras.
Harms
Over five million numerical facial representations were captured from shoppers at 12 Canadian malls without their knowledge or consent, and a third-party technology provider retained the biometric data on a decommissioned server.
Undisclosed facial recognition cameras were embedded in digital directory kiosks to analyze shoppers' age and gender without any signage or disclosure, capturing data from visitors across five provinces.
Affected Populations
- shoppers at 12 Canadian malls
- privacy rights advocates
Entities Involved
Embedded facial recognition cameras in digital directory kiosks at 12 shopping malls across five provinces, capturing over five million facial representations without shopper knowledge or consent
Led joint investigation with Alberta and BC privacy commissioners, finding Cadillac Fairview violated PIPEDA by collecting sensitive biometric information without meaningful consent or transparency
Responses & Outcomes
Published joint investigation finding that Cadillac Fairview violated PIPEDA by collecting biometric information without meaningful consent, and recommended deletion of data and express consent for any future use
Confirmed deletion of collected biometric data and removal of facial recognition cameras from mall kiosks
AI System Context
Facial recognition software embedded in digital directory kiosks at 12 Canadian shopping malls operated by Cadillac Fairview. The system captured shopper images and used facial analysis to estimate age and gender without knowledge or consent.
Preventive Measures
- Prohibit deployment of facial recognition systems in retail environments without explicit, informed consent from individuals whose biometric data is captured
- Require privacy impact assessments before any deployment of biometric surveillance technology in spaces accessible to the public
- Mandate that businesses disclose the use of facial recognition and biometric analysis through clear, prominent signage
- Establish enforcement mechanisms with meaningful financial penalties for undisclosed biometric data collection
Related Records
- clearview-rcmp-facial-recognition related
- canadian-tire-facial-recognition related
- Canadian Tire Deployed Facial Recognition to Identify Shoppers in British Columbia Stores related
- Facial Detection Cameras in Digital Ads Near Toronto's Union Station Scanned Commuters Without Consent for Three Years related
Taxonomy
Sources
AIID: Incident #358
Changelog
| Version | Date | Change |
|---|---|---|
| v1 | Mar 8, 2026 | Initial publication |