Algorithmic Cultivation: How Social Media Feeds Shape User Language

Collection of recent arXiv papers (May 19, 2026) with multiple security-relevant items. Notable works: MV-Gate — a new multi-view insider-threat detection framework that fuses statistical recurrence/frequency signals with tokenized activity sequences using an anomaly-aware gating mechanism; demonstrates improved detection on CERT r4.2/r5.2 and ADFA-LD, especially for low-signal progressive insiders. A study of algorithmic feed effects on language (Bluesky) shows sustained feed exposure causes measurable stylistic and psycholinguistic shifts, with implications for platform manipulation and user