mlsanityData to 7 Oct 2026

Recommender system in X inadvertently profiles ideological positions of users

arXiv 2602.02624 ↗First posted 2 Feb 2026cs.SI cs.AI cs.CYcited by 0

Not on the map. Why?

Abstract (arXiv)

Several data protection laws restrict processing that reveals political opinions, irrespective of the controller's intent. Whether recommender systems do so as a by-product of optimizing relevance has not been measured. From 2.5 million ``Who to Follow'' recommendations shown to 682 volunteers in France, we reconstructed an approximation of the embedding used by X's recommender for 26,509 accounts, computing survey-calibrated ideology scores. One direction in this embedding orders users by Left-Right position (Pearson rho = 0.887), distinct from directions tracking age, gender or popularity. We show this scale exists and affects the recommendations computed from the embedding. Removing it diversified recommendations at a limited cost in accuracy. We document a consequential form of emergent political representation that current definitions of profiling do not clearly address.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 16 of its 53 references that are in the map. First posted on arXiv 2 Feb 2026.

Builds on

16 of 53 references.

References · 16

  1. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,462
  2. 2020Open Graph Benchmark: Datasets for Machine Learning on Graphscited by 2,148
  3. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
  4. 2022Toy Models of Superpositioncited by 928
  5. 2020Null It Out: Guarding Protected Attributes by Iterative Nullspace Projectioncited by 442
  6. 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Themcited by 268
  7. 2019PyTorch-BigGraph: A Large-scale Graph Embedding Systemcited by 118
  8. 2022TwHIN-BERT: A Socially-Enriched Pre-trained Language Model for Multilingual Tweet Representations at Twittercited by 44
  9. 2021Algorithmic Amplification of Politics on Twittercited by 38
  10. 2023Engagement, User Satisfaction, and the Amplification of Divisive Content on Social Mediacited by 30
  11. 2022TwHIN: Embedding the Twitter Heterogeneous Information Network for Personalized Recommendationcited by 25
  12. 2023Embedding Democratic Values into Social Media AIs via Societal Objective Functionscited by 25
4 more
  1. 2019Demographic Inference and Representative Population Estimates from Multilingual Social Media Datacited by 24
  2. 2022Twitter Topic Classificationcited by 22
  3. 2022Fine-Grained Prediction of Political Leaning on Social Media with Unsupervised Deep Learningcited by 8
  4. 2024Echo Chambers in the Age of Algorithms: An Audit of Twitter's Friend Recommender Systemcited by 3

Built on by

Not cited yet.