← Back to catalogue/Federal (Canada)PIPEDA Findings #2020-004
Federal (Canada)Personal Information Protection and Electronic Documents ActWell-founded & resolved

PIPEDA Findings #2020-004: 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

Organization: The Cadillac Fairview Corporation LimitedComplainant: Anonymous applicant
Plain-language brief

The Office of the Privacy Commissioner of Canada (OPC), along with its provincial counterparts in Alberta and British Columbia, conducted a joint investigation into The Cadillac Fairview Corporation Limited (CFCL) regarding its use of Anonymous Video Analytics (AVA) technology in mall directories and mobile device geolocation tracking. For the AVA technology, the Offices found that CFCL collected and used personal information, including sensitive biometric numerical representations of faces, without valid consent. CFCL also improperly retained approximately 5 million such representations and video/audio recordings. The Offices concluded that CFCL contravened PIPEDA and provincial privacy acts regarding consent and retention for AVA. In response, CFCL ceased using the AVA technology, deleted the improperly retained data, and committed to staff training, leading to a "well-founded and resolved" outcome for this issue. For mobile device geolocation tracking, the Offices found that data collected from anonymous shoppers (hashed MAC addresses and non-granular zone geolocation) did not constitute personal information. Furthermore, CFCL clarified that geolocation data was not linked to identifiable logged-in Wi-Fi users. Consequently, this aspect of the complaint was deemed "not well-founded." The Offices, however, recommended that CFCL obtain express consent if it were to activate geolocation tracking for identifiable Wi-Fi users in the future.

Key issues
  1. 1Whether CFCL’s use of Anonymous Video Analytics (AVA) technology, via in-mall directories, resulted in the collection, use, and/or disclosure of personal information.
  2. 2Whether images of individual faces captured by AVA technology constitute personal information.
  3. 3Whether numerical representations of faces (biometric information) generated by AVA technology constitute personal information.
  4. 4Whether age range and gender assessments, combined with other data, constitute personal information.
  5. 5Whether CFCL obtained adequate and meaningful consent for the collection, use, and/or disclosure of personal information via AVA technology.
  6. 6Whether CFCL retained personal information collected via AVA technology longer than necessary.
  7. 7Whether CFCL’s use of mobile device geolocation technologies (Anonymous Shopper Journey) resulted in the collection, use, and/or disclosure of personal information.
  8. 8Whether hashed and randomized MAC addresses, combined with non-granular zone geolocation, constitute personal information in the context of anonymous shopper tracking.
  9. 9Whether CFCL’s use of mobile device geolocation technologies (Logged In Shopper Journey) resulted in the collection, use, and/or disclosure of personal information linked to identifiable individuals.
  10. 10Whether CFCL obtained adequate and meaningful consent for the collection, use, and/or disclosure of personal information via mobile device geolocation technologies (Logged In Shopper Journey).
  11. 11Whether CFCL's privacy policy and signage provided sufficient notice and obtained valid consent for its data collection practices.
  12. 12Whether the "serious possibility" threshold for identifying individuals was met for anonymous shopper journey data.
Outcome breakdownFavours: Both, in part
  • Consent for AVA technology: Contravention of privacy acts found
  • Retention of AVA data: Improper retention found
  • Resolution of AVA issues: Issues resolved by CFCL actions
  • Geolocation data as personal info: Anonymous geolocation not personal information
  • Geolocation tracking complaint: Complaint not well-founded
  • Future geolocation consent: Recommendation for express consent made
Outcome

AVA technology practices well-founded and resolved; geolocation tracking practices not well-founded.

Reasoning

CFCL collected sensitive biometric information via AVA technology without valid consent and retained it unnecessarily, but took corrective actions. For geolocation, the data collected from anonymous shoppers was not personal information, and for logged-in shoppers, geolocation was not linked to identifiable accounts.

AI-generated summary for reference only. Always verify against the official decision ↗
Decision notes
Recommended action / remedy

CFCL ceased use of the AVA technology, deleted the 5,061,324 numerical representations of faces and audio/video recordings, and provided privacy-related training to guest services employees. CFCL also removed language about 'special location-based offers' from its privacy policy and clarified what limited location information is collected.

Statutes considered
  • Principle 4.3 PIPEDA
  • section 6.1 PIPEDA
  • Principle 4.5.3 PIPEDA
  • section 7(1) PIPA AB
  • section 13(1) PIPA AB
  • section 35 PIPA AB
  • section 6 PIPA BC
  • section 7(1) PIPA BC
  • section 10(1) PIPA BC
  • section 35 PIPA BC

This summary is informational only and not legal advice.

Pro · AI

Decisions with similar reasoning and facts — found by AI across statutes and jurisdictions, not just keywords.

Pro
Coverage — 13 of 14 jurisdictions searchable

Fully searchable: Ontario, British Columbia, Alberta, Saskatchewan, Newfoundland and Labrador, Northwest Territories, New Brunswick, Manitoba.

Partial (recent decisions only): Federal (Canada) (651 of 1,631), Nova Scotia (49 of 472), Quebec (6 of 7,090), Prince Edward Island (2 of 279), Yukon (1 of 75).

Coming soon: Nunavut — being re-processed for AI search.

Find decisions like this one — by meaning, not keywords.

Related by meaning is a Pro feature. Upgrade to surface conceptually similar decisions across the corpus, powered by AI.

Upgrade to Pro