Tiered Anonymity on Social-Media Platforms as a Countermeasure against Deepfakes and LLM-Driven Mass Misinformation

TL;DR

The paper addresses the problem that AI-generated deepfakes and rapid LLM-driven misinformation exploit the asymmetry of online influence by enabling broad reach under blanket anonymity. It proposes a reach-based, three-tier anonymity framework where a user's reach score $R$ determines escalating duties: Tier 1 preserves full pseudonymity, Tier 2 requires platform-held private identity linkage, and Tier 3 imposes independent, ML-assisted fact-checking and provenance controls before wide distribution. The authors contribute a formal model mapping $R$ to identity obligations, an empirical Reddit case study showing proportional, endogenously emergent moderation gates, and a regulatory pathway grounded in the EU's DSA and the UK's Online Safety Act to operationalize these controls. The framework aims to restore friction and accountability to influential online voices while preserving privacy for ordinary users, aligning platform incentives with democratic values and reflecting evolving international regulatory trajectories.

Abstract

We argue that governments should mandate a three-tier anonymity framework on social-media platforms as a reactionary measure prompted by the ease-of-production of deepfakes and large-language-model-driven misinformation. The tiers are determined by a given user's $\textit{reach score}$: Tier 1 permits full pseudonymity for smaller accounts, preserving everyday privacy; Tier 2 requires private legal-identity linkage for accounts with some influence, reinstating real-world accountability at moderate reach; Tier 3 would require per-post, independent, ML-assisted fact-checking, review for accounts that would traditionally be classed as sources-of-mass-information. An analysis of Reddit shows volunteer moderators converge on comparable gates as audience size increases - karma thresholds, approval queues, and identity proofs - demonstrating operational feasibility and social legitimacy. Acknowledging that existing engagement incentives deter voluntary adoption, we outline a regulatory pathway that adapts existing US jurisprudence and recent EU-UK safety statutes to embed reach-proportional identity checks into existing platform tooling, thereby curbing large-scale misinformation while preserving everyday privacy.