Furthering our work in uncovering what the Canada Border Services Agency (“CBSA”) is doing, I am sharing the draft Algorithmic Impact Assessment we have for their Security Screening Automation (“SSA”) project.
Based on a recent information technology report (disclosing projects with a budget of more than $1 Million dollars), the following is summarized about this automation project:
Security Screening Automation
The security screening system will facilitate information sharing with screening partners and enhance security screening through a new case management system, integrated intelligence systems and tools, business intelligence and performance monitoring capabilities.
We also learn from this document that completion is tabled for 2026.
From CBSA’s 2024-2025 annual report, we also see there are five references to the SSA project suggesting it is a major departmental priority.
The details are still quite mum about the project stating only, their goal to:
Enhance immigration security screening processes to identify inadmissible persons seeking
entry into Canada, including efforts under the Security Screening Automation project
Algorithmic Impact Assessment
Much of this document (which I received last year in 2023) is redacted. We do not know when the draft was dated.
The project description (Section 9 at page 1) describes that it will invovled modernized case management system, an automated triage function, integrated intelligence systems, and tools, business intellignce, and performance moniotring capabilities. In short, this looks much more advanced then a Chinook type IRCC model.
It also re-affirms the use of htese tools to route screening for the easier cases and allot more time for review for complex cases – helping hte immigration client. The document reinforces again that no automated decision is made without human oversight. The actual explanation for “Human-in-the-loop for decisions” is redacted.
I wonder how automation of straight forward cases for security will be received by the public in light of recent news around terrorist plots by individuals who apparently were not caught by our security screening apparatus. It also seems somewhat contradictory that CBSA appears (again: I acknowledge I’ve received this information only through hearsay) concerned about IRCC’s eligiblity screening processes being automated, while pursuing security screening automation on the other hand.
Another question I have is whether and how facial recognition technology (“FRT”) being implemented. We know that FRT is implemented into a ReportIn tool being developed to facilitate the removals process for inadmissible individuals subject to removal. Presumably, in addition to the obvious analytics from the application materials an individual might have and require processing, there is also a photo as part of every application and biometric collection. I wonder if there is any role being played in analyzing this as part of a front-end screening.
Anyways, I leave the document up for everyone’s review. Happy to have further discussion about this:
Interim Release A-2023-18296For those who cannot see the above document, you can access the document via Google Drive link here: https://drive.google.com/file/d/1OBBoWuYWysvdyD1b2dG4B6E3m7Z0hylq/view?usp=drive_link
