Algorithmic curation and deliberative democracy: A critical analysis of echo chambers, platform governance, and digital public discourse

Abstract

Scholarship on social media and deliberative democracy has converged on a single, repeated claim since roughly 2015: that social media both enables and undermines democratic discourse. This claim no longer advances the field, because it does not specify which architectural feature of a platform produces which deliberative outcome, and because it rarely tests its own assumptions against direct evidence of platform governance decisions. This study isolates algorithmic curation, as distinct from social media as a communication medium, as the primary object of analysis, and asks what effect it has on deliberative democracy, why that effect occurs, and how it can be governed. The study conducts a qualitative thematic synthesis of fifty-two English-language secondary sources, drawn from an initial pool of eighty-three records identified through structured searches of Google Scholar, JSTOR, SSRN, and Scopus-indexed journals, screened against explicit inclusion and exclusion criteria, and coded against four deliberative norms: inclusiveness, reasoned argumentation, reciprocity, and reflexivity. The coded literature is read against four documented cases of platform governance: Meta’s January 2025 termination of third-party fact-checking, the measured limits of X’s Community Notes system, the structural fragility of Mastodon’s federated moderation model, and Nepal’s September 2025 Gen Z movement, in which a government ban on twenty-six platforms triggered mass protest and a subsequent selection of an interim prime minister conducted partly through a hundred-thousand-member Discord server.  Social media expands the formal opportunity to participate while algorithmic amplification reproduces existing inequalities of voice; engagement-optimized ranking, not social media as a medium, is the mechanism most responsible for the decline of reasoned and reciprocal exchange; and current institutional responses, crowd-sourced correction, federated decentralization, and crisis-driven hybrid deliberation, each restore some deliberative capacity but none restores it fully. Deliberative democratic theory needs a distinct, empirically grounded account of algorithmic governance, built on a broader evidentiary base than the Western regulatory literature it has so far relied on.