AI and Administrative Law
in Florian Martin-Bariteau & Teresa Scassa, eds., Artificial Intelligence and the Law in Canada (Toronto: LexisNexis Canada, 2021)
2 Pages Posted: 4 Dec 2020
Date Written: November 20, 2020
Algorithmically-driven decision-making tools have long existed in Canadian administrative agencies, but administrative law largely fails to address such technologies. This chapter uses the basic principles enunciated in administrative law doctrines to demonstrate how algorithmically-driven decisions fundamentally challenge administrative law notions of fairness, responsibility, and justice. In doing so, it identifies which common law principles, such as notice requirements or prohibitions on sub-delegation, might best apply to algorithmic tools. Other doctrines, it argues, must adapt to meaningfully address the problems algorithmic technologies raise for procedural fairness, responsible and responsive decision makers, and substantive justice. A central dilemma, and an area for future evolution, is how administrative law ought to conceptualize algorithmic tools: should it approach them as a forceful policy, a uniquely persuasive source of evidence, or as though they are an additional decision maker? This chapter illustrates how different answers to this question may require varied doctrinal shifts. It then pinpoints opportunities for legislators and policymakers to develop new governance frameworks that address the dilemmas of algorithmically-driven decision-making. When crafting statutes, policies, and institutional arrangements, legislators and policymakers must meaningfully address the interests of marginalized communities, as algorithmically-driven tools tend to impact members of these communities most harshly. This proposal marks a new path for this area of law, as administrative law principles typically protect individual rights and interests but are less adept at confronting systemic injustice.
Keywords: AI; administrative law; government; automated decision-making; Algorithmically-driven decision-making tools
Suggested Citation: Suggested Citation