The study of how online platforms, digital communities, and algorithmic systems facilitate the spread of extremist ideologies, enable coordinated extremist activity, and contribute to individual radicalization. This field examines both the mechanisms of online extremism and potential interventions.
Key research questions
- Platform role in radicalization: Do social media algorithms, recommendation systems, and content moderation choices drive ideological radicalization, or do they reflect user demand for extremist content?
- Radicalization pathways: What are the primary routes by which users encounter extremist content—algorithmic recommendations, subscriptions, external links, search, or direct community participation?
- Attitudinal factors: What prior attitudes (resentment, grievance, in-group loyalty) correlate with or predict consumption of extremist content?
- Concentration and influence: Is extremist viewership concentrated among a small, dedicated audience, or does it reach mass audiences?
- Intervention effectiveness: Can platform policies, content removal, de-amplification, or labeling reduce extremist reach without suppressing legitimate political speech?
Key papers
- Chen et al. (2023) — Subscriptions and external links help drive resentful users to alternative and extremist YouTube channels — Empirical study using browser tracking showing that exposure to alternative and extremist YouTube channels is driven primarily by user demand (high gender and racial resentment) and deliberate seeking (subscriptions and external links), not algorithmic recommendations. Finds algorithmic "rabbit holes" extremely rare (0.01% of visits). YouTube's 2019 moderation changes appear to have reduced unwanted algorithmic recommendations.
- Hosseinmardi et al. (2021) — Large-scale study of 309K US YouTube users over 4 years; finds little evidence of systematic algorithmic radicalization to far-right content; far-right audience remains small and non-growing; anti-woke content grew 3× but correlates with off-platform far-right consumption, suggesting user preference rather than algorithmic gateway effect; 55% of far-right videos accessed via external URLs or search, not recommendations.
- Horta Ribeiro et al. (2019) — Auditing Radicalization Pathways on YouTube — First large-scale audit of radicalization on YouTube; tracks 349 channels across Intellectual Dark Web, Alt-lite, and Alt-right; uses random walk simulations to show Alt-right channels are ~50% reachable from milder ideological communities via YouTube recommendations; provides evidence for "rabbit hole" theory but does not measure actual user behavior.
Related concepts
- youtube-radicalization — YouTube-specific radicalization dynamics
- platform-algorithms — Algorithmic design and amplification mechanisms
- content-moderation — Platform policies and enforcement
- computational-propaganda — Organized extremist campaigns
- social-bots — Automated accounts amplifying extremist content