Canada’s proposed Artificial Intelligence and Data Act (AIDA) was built upon a fundamental lack of democratic input. It is a hallmark example of AI regulation driven by a narrow, expert-led process, a problematic approach given its pronounced democratic deficit. Without input from many marginalized and high-impact communities, including civil society organizations, Indigenous governments, disability advocates, labour groups, racialized communities and youth, the bill was constructed on sand from the start.
This could provide an important lesson for Canada and other democracies at this pivotal moment in AI regulation, however. There must be a new understanding of what is needed to craft democratic AI governance that is inclusive and reflective of the interests of those most vulnerable to algorithmic harms. Julian Lam provides actionable recommendations for policy makers to put these lessons into motion.
