An Analysis of the Artificial Intelligence and Data Act: Envisioning Feminist Alternatives for Mitigating Algorithmic Harm

Candidate: Laine McCrory, Communucation and Culture PhD

This thesis conducts an in-depth policy analysis of how feminist theory and policy can be used to conceptualize and mitigate AI harm through a collective, justice-based approach. Using a research framework that blends reflexive thematic analysis (Braun and Clarke, 2006) and situational analysis (Clarke, 2005), this research analyzes a sample of 86 documents related to the Canadian Artificial Intelligence and Data Act. The research established three central themes related to the role of 1) diverse perspectives influencing what constitutes AI harm; 2) equity-focussed policy in addressing AI Harm; and 3) public engagement and oversight within the policymaking process. This thesis concludes with an analysis of these themes through a forward-looking feminist lens to argue that a feminist AI policy should see harm as connected to social structures and systems of power to be addressed through the cultivation of resistant knowledge, a situated ethics of collectivity, and a relational accountability structure. Feminism; artificial intelligence; digital policy; feminist science and technology studies