1. How Recommendation Algorithms Run the World
Author: Zeynep Tufekci Year: 2019 Type: Journalism Publisher: WIRED URL: https://www.wired.com/story/how-recommendation-algorithms-run-the-world/
What this source contributes
A journalistic overview of recommendation logic and its societal consequences, giving an accessible framing of the engagement-optimisation problem.
Related entries
- Recommender Systems — public-facing framing of the optimisation problem
2. YouTube, the Great Radicalizer
Author: Zeynep Tufekci Year: 2018 Type: Op-ed Publisher: The New York Times URL: https://www.nytimes.com/2018/03/10/opinion/sunday/youtube-politics-radical.html
What this source contributes
The op-ed that named the radicalization pathway, widely cited as the public entry point to the debate and predating the empirical studies that tested the claim.
Related entries
- Recommender Systems — the naming of the pathway before it was measured
3. Algorithmic extremism: Examining YouTube’s rabbit hole of radicalization
Author: M. Ledwich & A. Zaitsev Year: 2019 Type: Academic article (contested) Publisher: First Monday URL: https://www.researchgate.net/publication/339634318_Algorithmic_extremism_Examining_YouTube’s_rabbit_hole_of_radicalization
What this source contributes
A contested empirical study arguing that YouTube’s algorithm actually favours mainstream and left-leaning channels over extremist ones. Included as a counterpoint to the radicalization-pathway consensus, and worth keeping precisely because the entry’s other sources agree with each other.
Related entries
- Recommender Systems — the dissenting reading of the same platform
4. Emotion shapes the diffusion of moralized content in social networks
Author: W.J. Brady, J.A. Wills, J.T. Jost, J.A. Tucker & J.J. Van Bavel Year: 2017 Type: Academic article Publisher: PNAS URL: https://doi.org/10.1073/pnas.1618923114
What this source contributes
Empirical study showing that moral-emotional language spreads furthest on social platforms. Supports the anger and fear engagement-signal claim in the Friction section. Serves as a primary source in the Ragebaiting entry, where the mechanism itself is the object of analysis.
Related entries
- Recommender Systems — the emotional signal the system optimises against
- Ragebaiting — the same finding used there as a primary source
5. A Survey of Affective Recommender Systems
Author: Kang et al. Year: 2025 Type: Academic survey Publisher: arXiv URL: https://arxiv.org/html/2508.20289v1
What this source contributes
A systematic survey of how emotions are modelled as parameters in recommender systems, covering Ekman’s basic emotions, valence-arousal models, and Plutchik’s circumplex. Theoretical background for the anger and fear engagement mechanism, and evidence that emotional response is an explicit design variable rather than a side effect.
Related entries
- Recommender Systems — emotion as a modelled parameter in system design
6. Hope, Fear, or Anger? How Emotional Framing in a News Recommender System Influences User Engagement and Openness to Non-Preferred Content
Author: author to be confirmed Year: 2023 Type: Conference paper Publisher: CEUR Workshop Proceedings URL: https://ceur-ws.org/Vol-4056/full5.pdf
What this source contributes
User study (N=150) finding that anger and fear are high-arousal emotions that capture attention and increase engagement in news recommender systems. Direct empirical support for the anger and fear mechanism.
Related entries
- Recommender Systems — experimental support for the arousal mechanism
- Filter Bubble — openness to non-preferred content as the variable under test
7. Big Tech maakt van burgers vee
Author: Roxane van Iperen Year: 2026 Type: Journalism / essay Publisher: Vrij Nederland, 9 July 2026 URL: https://www.vn.nl/op-afstand-gestuurde-mens
What this source contributes
The image the entry borrows for the whole logic: feed, notification and recommendation algorithm as a virtual fence, which does not merely steer users through the available content but decides how much they consume and which territories exist for them at all. A Dutch essay rather than research; its value is the metaphor, and the fact that the argument is made for a general public in Dutch. Also cited in Filter Bubble, Dopamine Feedback Loops, Biometric Surveillance and Surveillance Capitalism.
Related entries
- Recommender Systems — the fence, in Friction
- Filter Bubble — the same essay on the bubble as a steered environment