Consider these dilemmas: A hospital's AI triage system consistently prioritizes younger patients over elderly ones because its training data shows younger patients have better outcomes. Is this fair or ageist? A school uses AI to flag students at risk of dropping out, but the system disproportionately flags students from low-income families, potentially creating a stigma. A self-driving car must choose between swerving into a barrier (risking its passenger) or staying course (risking a pedestrian). Who decides the algorithm's priorities? A content moderation AI removes a human rights documentation video because it contains graphic violence. An AI tutoring system adapts its teaching based on predicted student ability, potentially limiting what some students are exposed to. There are no clean answers here, only trade-offs.