Canadian ai governance reveals power struggles and hidden voices
Infrastructure and its Discontents: Contestations of Power in Canada's AI Governance Scaffold
Human-Computer InteractionComputers and Society
Summary
Governments try to manage how artificial intelligence is used, but these efforts can hide some voices and favor others. The authors studied how Canada's AI rules are made and found that experts’ power and unclear information often cause problems. They also saw that current systems struggle to fix harms caused by AI. They suggest new ways to design AI governance that focus on fairness, care, and including more people in decisions.
What this means in practice
- •For policy makers: Guide the design of AI policies that include diverse voices and address power imbalances in governance structures.
- •For civil rights advocates: Support advocacy strategies focused on demanding transparency and accountability from AI governance bodies.
A position paper. It proposes an approach and reports no results.
Authors
Christelle Tessono, Nusrat Jahan Mim, Shion Guha
Abstract
This paper develops a critical theoretical framework for understanding national Artificial Intelligence (AI) strategies through the lens of "Governance as Infrastructure." Integrating insights from Infrastructure Studies, Science and Technology Studies (STS), this paper employs Black Feminist Technoscience (BFTS) to reveal how AI governance infrastructures are assembled, maintained, and contested, and whose forms of knowledge, labour, and experience are rendered visible or invisible in these processes. We analyze the Canadian AI governance landscape through in-depth interviews with bureaucrats, civil society advocacy groups, academic researchers, and industry actors. Our inductive thematic analysis reveals how AI governance infrastructure enacts specific configurations of power, particularly through the politics of expertise; how actors navigate profound ambiguities related to AI's opacity, often involving strategic ignorance; and how mechanisms for justice and accountability struggle to address AI-related harms. We contend that a BFTS perspective highlights the limitations of current governance approaches and points toward alternative infrastructural designs prioritizing care, maintenance, participatory democracy, and adaptive temporalities to foster more equitable and accountable technology governance.