Contesting the public interest in AI governance

Abstract

This article argues that public contestability is a critical attribute of governance arrangements designed to align AI deployment with the public interest. Mechanisms to collectively contest decisions which do not track public interests are an important guardrail against erroneous, exclusionary, and arbitrary decision-making. On that basis, we suggest that efforts to align AI to the public interest through democratic participation will benefit substantially from strengthening capabilities for public contestation outside aggregative and deliberative processes. We draw on insights from democratic and regulatory theory to explore three underlying requirements for public contestability in AI governance: (1) capabilities to organise; (2) separation of powers; and (3) access to alternative and independent information. While recognising that suitable mechanisms for contestability will vary by system and context, we sketch out some possibilities for embedding public contestability in AI governance frameworks with a view to provoking further discussion on institutional design.