情境
A user watches a video and the YouTube recommendation algorithm logs engagement; if fringe content dominates, repeated boosts and profile shifts can push the user toward a radicalized worldview.
這張圖裡有什麼
看懂背後的決策。
01
Does the recommendation engine score mainstream or fringe affinity?
02
Does the user engage with more fringe content after the feed is dominated?
03
Has repeated exposure crossed the radicalization risk threshold?
Use this flowchart to explain or audit recommender-system risks for extremism, train content moderators, or design deranking interventions.