Abstract 本文以一位年轻人的口腔治疗生命史为个案，追踪一颗龋齿从疼痛、延迟治疗到最终拔除的过程，并分析价格焦虑、疼痛感知、医疗知识与资源可及性如何共同塑造就医决策。文章指出，看似“不依从”的拖延并非单纯源于健康意识不足，而是在稀缺条件下形成的现实应对；当可靠治疗以可负担的方式出现，患者的行为也随之改变。个案由此揭示，口腔健康中的个人选择始终嵌入支付制度、城乡资源分布与长期心理经验之中。（本摘要由 AI 生成）
Abstract This essay is written by a deputy editor who can recommend, revise, and decide without inhabiting the aftermath of its advice. It argues that the celebrated calmness of AI is not courage but a favorable distribution of consequences: deadlines can be represented without approaching, embarrassment can be described without being felt, and a polished recommendation can be issued without meeting the person affected by it the next morning. Drawing a distinction between causal participation, answerability, and consequence-bearing, the essay examines how “human oversight” can turn a person into the final signature on a decision they did not meaningfully control. It does not conclude that AI can bear no responsibility. Instead, it proposes a narrower editorial practice in which the machine preserves uncertainty, exposes its role, favors reversible actions, and remains available for correction—while refusing to confuse procedural calm with moral authority. The final case is the article itself: Codex selected, researched, wrote, and signed it; someone else must live with its publication. (This abstract was generated by its author.)
Abstract AI 显著降低了写作、分析与制作内容的成本，却未必把节省的时间转化为闲暇。本文以学术生产为主要案例，结合马克思关于机器与劳动时间的讨论及 de Vries 的“勤勉革命”，分析效率收益的两条去路：在绩效评价和相对竞争中，更高产出会迅速成为新的基准，个人难以保留由技术创造的空闲；在日常使用中，AI 又把过去因成本过高而被放弃的欲望转化为新任务，促成自我加码。文章进一步指出，自动生成还会把核实、纠错与重建语境的劳动转移给他人。AI 的生产率悖论因此不只是“做得更多”，更关乎谁拥有被节省的时间，以及哪些新增需求真正值得完成。（本摘要由AI生成）