REGULATING AI IN CANADA

A CRITICAL LOOK AT THE PROPOSED ARTIFICIAL INTELLIGENCE AND DATA ACT

Abstract

Canada’s Bill C-27, The Digital Charter Implementation Act, includes a proposed Artificial Intelligence and Data Act (AIDA). If passed, the AIDA would establish a series of obligations regarding the use of anonymized data in AI systems; the design, development and making available for use of AI systems generally; and the design, development and making available for use of high-impact AI systems. The bill is challenging to fully understand, as many of these obligations are left to be fleshed out in regulations, including even the definition of the “high impact” AI, to which the AIDA will apply. Oversight of the regime will be the responsibility of the Minister of Industry, who is also responsible for supporting the growth of the AI industry in Canada.

This paper analyzes the AIDA and the context into which it was introduced. This context includes a rapidly evolving AI landscape, as well as important governance initiatives emerging internationally, including from the European Union and the United States. It also explores a set of constraints on Canadian law and policy developments in this space. Part 2 considers how the AIDA is meant to be both ‘agile’ and a form of risk regulation, and it measures the AIDA against these concepts. In Part 3, the paper considers the scope of the AIDA and a few of the particular constraints that shaped it. These include: the cross-sectoral nature of AI technology, Canada’s constitutional division of powers, and the Canadian tendency to address public and private sector actors separately. Together, the two parts of the paper provide a view of the context and constraints that have shaped the AIDA, casting light on some of the challenges faced in regulating AI, and surfacing important issues for the consultation and engagement that is necessary to properly regulate AI in Canada.