Is Artificial Intelligence Really Undermining Democracy? A Critical Appraisal of Mark Coeckelbergh’s Why AI Undermines Democracy (2024)

Mark Coeckelbergh’s recent book, Why AI Undermines Democracy (2024), argues compellingly that artificial intelligence (AI) intrinsically threatens democratic governance by facilitating epistemic manipulation, deepening knowledge asymmetries, and fostering political alienation. While acknowledging Coeckelbergh’s valuable insights into democratic vulnerabilities exacerbated by digital technologies, this essay critiques his foundational assumption that AI inherently possesses political or epistemic agency. Drawing upon fiduciary epistemic theory—particularly elaborated in my works ‘Epistemic Justice and Institutional Responsibility in Academia’ (205) and “Directors’ Epistemic Duties and Fiduciary Openness” (2025)—I propose an alternative governance framework, termed ‘epistemocracy’, which explicitly locates epistemic and moral responsibility within human institutional actors rather than technological artefacts. Unlike Coeckelbergh’s ‘technodemocracy’, epistemocracy prioritises fiduciary transparency, institutional accountability, and epistemic pluralism as foundational to democratic governance of AI. Ultimately, the essay argues that attributing inherent political agency to AI obscures crucial institutional responsibilities and weakens democratic safeguards. Recognising AI as epistemically neutral and embedding fiduciary epistemic duties into governance structures provides clearer, more robust pathways for institutional reform, democratic accountability, and sustainable epistemic justice in AI governance.