Open Collections

Abstract

As a result of fast-moving developments in artificial intelligence (“AI”) tools such as machine-learning (“ML”), Canadian scholars have been engaged in a recent, concerted effort to examine how the use of automated (often also termed “algorithmic”) decision-making systems (“ADMs”) by public administration officials may alter traditional decision-making processes. This examination has been focused on exploring the impacts these technologies are having on foundational administrative law principles, primarily through the lens of ex post adjudication and judicial review. This thesis continues this line of scholarship by exploring a specific problematic, how the use of ADMs by Immigration, Refugees and Citizenship Canada (“IRCC”), is altering the process of decision-making in a way that necessarily creates implications for the way external mechanisms like judicial review, are able to aid in reviewing decisions. This enquiry is driven through doctrinal method, applying a law and technology approach, and applying Michael Adler’s administrative justice theories and typologies. Tracing how decisions have shifted, I argue that failures to understand process are impacting both procedural fairness and reasonableness review. Furthermore, I argue that the ability to understand and interrogate process, as a prerequisite, requires greater transparency, accountability, and structures of ex ante rulemaking. I conclude with a strong recommendation for structural reform aimed at “getting it right the first time” – suggesting a starting point of procedural protections within statute, namely the Immigration and Refugee Protection Act and developing procedural code, inviting refinement.