Platform enforcement refers to how social media companies and online services apply their content policies in practice—including the mechanisms (suspension, deplatforming, warnings, labeling, algorithmic suppression), the decision-making processes, and the outcomes of enforcement actions.
Key aspects
Enforcement mechanisms: Platforms deploy multiple enforcement tools including account suspension or permanent bans, content removal, warning labels, fact-check annotations, algorithmic de-amplification, and rate-limiting. Each mechanism has different efficacy and user-impact profiles.
Policy-practice gap: The relationship between stated policies (which are often politically neutral) and actual enforcement outcomes can be complex. Asymmetries in enforcement outcomes may reflect asymmetries in behavior, policy design, implementation bias, or some combination.
Efficacy and spillovers: Research examines whether enforcement actions successfully reduce misinformation spread and whether they have unintended consequences (e.g., driving users to alternative platforms, triggering backlash).
Equity and transparency: Critical questions include whether enforcement is applied fairly across groups, whether users understand why they were sanctioned, and whether appeals processes exist and function effectively.
Key papers
- Mosleh et al. (2024) — Differences in misinformation sharing can lead to politically asymmetric sanctions — Shows that political asymmetries in platform suspension rates can arise from asymmetric misinformation-sharing patterns, without requiring intentional bias in policy enforcement.
- McCabe et al. (2024) — Twitter's mass deplatforming following January 6 — Evaluates the effectiveness of platform enforcement through deplatforming; uses natural experimental designs to show the intervention reduced misinformation circulation among both deplatformed users and their followers.
- Bak-Coleman et al. (2022) — Combining interventions — Quantitative framework for evaluating individual and combined enforcement and correction interventions; shows combined approaches are substantially more effective than single interventions.
- Donovan (2020) — Mitigating Medical Misinformation During an Infodemic — Discusses platform enforcement responses to misinformation (content removal, fact-check labels, user bans) and how bad actors adapt to evade moderation.