AI, Deskilling, and the Prospects for Public Reason

Abstract The relationship between democracy and artificial intelligence (AI) is attracting attention, given fast-paced developments in AI and their implications for the political public sphere. The idea of ‘public reason’ can illuminate important dimensions of this relationship. Public reason is a standard of reciprocal legitimation and justificatory practice given democratic disagreement. This paper argues that AI might threaten the prospects of public reason when applied to policy debates. On the ‘civic friendship’ conception of public reason, the practice of reasoning is grounded in embodied joint action and having shared experiences. Accordingly, public reason as a reciprocal justificatory practice requires being skilled in human capacities like justice, joint action, patience and moral attention. Yet AI tools (e.g. recommender systems, personalized AI aids and AI deliberative democracy platforms) that afford disembodied, mediated interaction, threaten to deskill humans of those capacities, by creating environments that afford less opportunities to engage in activities in which the capacities are cultivated and practiced. AI tools involved in democratic deliberations can provide efficiency, scalability, and improved understanding of policy issues among participants. However, on balance, uncritical integration of such tools could deskill public reason capacities, leading to the erosion of mutual assurance between citizens, and ultimately undermining trust in democratic deliberation.