Facial Recognition

Date: July 3, 2020
Classification: Unclassified
Fully releasable (ATIP)? Yes
Branch / Agency: ISTB/CBSA

Proposed Response:

If pressed on racial profiling:

Background:

At the June 25th, 2020 Senate Committee of the Whole meeting which was held to discuss “Ending Systemic Racism towards Black people, Indigenous people and putting an end to systemic racism”, the Minister of Public Safety Bill Blair spoke to concerns being raised around the utilization of facial recognition (FR) and potential biases around software and racialized individuals. He was also queried on privacy protections surrounding FR. While the discussion focused largely on the Royal Canadian Mounted Police, the CBSA continues to explore the use of biometrics when confirming travellers identities, and the potential benefits in adopting FR procedures to facilitate legitimate travellers and detect individuals who pose a threat to the security of Canada. 

CBSA Current and Future Uses

The CBSA uses facial matching technology on the Primary Inspection Kiosk (PIK) at Canada’s largest 10 international air ports of entry, as well as on NEXUS kiosks at seven international airports. The CBSA began implementing PIK in 2017 in order to process increasing numbers of arriving travellers, as well as to address the public’s interest in increased self-service options.

The NEXUS program began a process of kiosk modernization in October 2019, building upon the PIK model and tailored specifically for NEXUS members. With PIK, the CBSA utilizes facial matching technology to support border processing activities on virtually all travellers arriving internationally, both returning Canadians and foreign nationals. Facial matching is not used for travellers under 14 years of age as well as for travellers who present travel document which do not contain an eChip (i.e. Canadian Permanent Resident card, passports of countries that have not migrated to newer electronic machine readable travel documents). In these cases, a Border Services Officer manually checks the traveller against the image displayed in the travel document. Since 2013, all new Canadian passports are issued with an eChip. Facial matching is not conducted when travellers are processed manually by an officer (e.g. when a traveller is not eligible or declines to use a PIK, or when travellers under the age of 16 travel unaccompanied by an adult).

Additionally, over the last few years, the CBSA has engaged in several applied research and development projects that have included evaluation of the readiness of facial recognition technology for border operations. The evaluation has been done through the development of two laboratory-based prototypes and one time-limited pilot in an operational environment. Each project was sponsored and partially funded by the (Canadian Safety and Security Program (CSSP)). Most specifically, the Faces on the Move (FoTM) project was conducted between August 2014 and March 2017. The project was sponsored by the CSSP and managed by the CBSA. The Face4 Systems Inc. was the industry partner leading the face recognition system deployment and performance assessment, the Adga Group lead the privacy impact assessment and the École de technologie supérieure (ETS) as the academic partner.

The CBSA continues to explore potential uses for facial recognition in establishing travellers identities. While this work is in it’s infancy, the CBSA is focusing on also ensuring the privacy of travellers, and exploring current concerns surrounding biases built into facial recognition technology. As the Agency continues this exploration, the Office of the Privacy Commissioner will be consulted to ensure that the CBSA is putting forward proposals that do not inherently discriminate against individuals. 

Facial Recognition versus Face Matching

Facial matching uses automated face systems to perform 1:1 verification (authentication) comparing an individual’s face to a registered face image to determine if the two faces are a match. This is typically used to verify the identity of an individual.

Facial recognition (FR) is an automated biometric systems for identification that employs a 1:N (many) search against a database of images to try to identify an individual. The automated system will compare the submitted image against a biometric database containing images of ‘known’ faces previously enrolled in the system, including in some cases watch list images. The search will result in the generation of a candidate list of images that meet the pre-established threshold for similarity. 

Contacts:

Approved by: Minh Doan, Vice-President, Information and Science Technology Branch, 613-948-9694

Date modified: