• CASES

    Search by

Doan v. Clearview AI Inc.

Executive Summary: Key Legal and Evidentiary Issues

  • Appeal focused on whether the class could be sufficiently identified under Rule 334.16(1)(b) of the Federal Courts Rules.

  • Ms. Doan challenged the lower court's conclusion that no workable method existed to identify class members.

  • The trial judge had rejected proposed query methods as creating an “opt-in” scheme, contrary to federal opt-out rules.

  • The Court of Appeal ruled that querying Clearview for confirmation did not convert the case into an opt-in proceeding.

  • Lack of full analysis by the certification judge on whether queries could identify class members during the opt-out period.

  • Appeal was allowed; the certification motion was remitted to the Federal Court for proper reconsideration.

 


 

Facts and outcome of the case

Background and parties
Ha Vi Doan, a Quebec-based photographer and Canadian citizen, brought a proposed class action against Clearview AI Inc., a U.S. company known for creating a vast database of facial images sourced from the internet. She alleged that Clearview infringed copyright and moral rights by collecting, storing, and indexing photographs that she and others had taken, many of which were publicly shared online without consent. Clearview’s technology allows law enforcement and other clients to perform facial recognition searches using this database.

Doan sought certification of her action as a class proceeding, claiming to represent a class of Canadian citizens or residents who are either authors or copyright holders of the collected photographs. She requested declaratory and injunctive relief, deletion of images from Clearview’s database, and damages.

Federal Court decision
The Federal Court rejected Doan’s motion to certify the proceeding as a class action. The judge found that Doan had not demonstrated a "some basis in fact" that an identifiable class existed, a requirement under Rule 334.16(1)(b) of the Federal Courts Rules. The judge also concluded that Doan’s proposed methods to identify class members—particularly through submitting queries to Clearview—would impermissibly convert the proceeding into an opt-in class action, contrary to the Federal Court's opt-out framework.

Issues on appeal
On appeal, the Federal Court of Appeal considered whether the certification judge erred in law and fact, particularly regarding the interpretation of the opt-out scheme and whether class members could be identified through proposed query mechanisms. The appellate court found the lower court had mischaracterized the query method as converting the case into an opt-in scheme and failed to fully assess whether the method could work during the opt-out period.

Outcome of the appeal
The Federal Court of Appeal allowed the appeal, ruling that the query method did not transform the class action into an opt-in process and that the certification judge erred by not thoroughly evaluating whether the method could help identify class members in time. The Court set aside the lower court’s decision and remitted the certification motion back to the Federal Court for reconsideration under the correct legal principles.

Costs and damages
No costs or damages were awarded in the appeal. The Court of Appeal specifically allowed the appeal without costs and did not assess any damages, as the matter was procedural and the underlying claims remain unresolved.

Clearview AI Inc.
Federal Court of Appeal
A-341-23
Intellectual property
Not specified/Unspecified
Appellant
11 December 2023