Answers are grounded in this paper and cite their sources.
Theoretical analysis uncovers mechanisms of labor-wealth decoupling in autonomous production, suggesting structural separation of human worth from capital market pricing.
# 摘要 / Abstract ## 中文 人工智能、机器人、智能代理、自动化基础设施与无人生产正在共同形成一种不同于过去技术升级的新型生产力。过去几次工业革命主要扩大了人的机械能力、能源利用能力、计算能力和信息传播能力,而无人化革命更深刻的潜在变化在于:越来越多的生产活动可以在不按比例增加人类有偿劳动的情况下完成。由此,真正需要重新提出的问题并不是简单的“AI会不会抢走工作”,而是:**当大规模人类劳动逐渐不再具有原有的生产必要性时,劳动是否还必须继续承担社会成员获得社会财富的主要分配中介功能?** 本文在三篇前置论文的基础上建立一个更完整的历史框架。第一层是“解构”:竞争性无人化对单个企业而言通常是理性的,但在普遍化后可能形成劳动收入减少、自动化收益集中与有效需求之间的结构性张力。若自动化收益集中于低消费倾向的资本所有者,而价格下降、新任务、投资、出口、财政转移与其他购买力来源不足以补偿劳动收入减少,那么私人自动化可能超过社会层面的最优水平。本文将其概括为“竞争性无人化的需求外部性”,并明确把它视为条件性、可证伪的机制命题,而非资本主义必然崩溃的定律。 第二层是“构建”:同一种无人化生产力也可能改变教育、医疗、照护、行政、交通和其他基础服务的成本结构。传统服务的生产者本身必须生活,住房、家庭、配偶就业、子女教育和区位偏好因此嵌入服务成本与地理分布。无人化并不意味着服务零成本,而可能使服务从“附着于必须在那里生活的人”转向“附着于可以部署的基础设施”。在此基础上,部分基本生活资料可能从“取得工资后进入市场购买”逐渐转向“作为社会成员直接接入”,从而形成“接入型社会化”。 第三层进一步提出一个比制度名称更深的问题:**如果一个人的社会生存越来越不需要他的劳动,那么资本还能否继续成为决定这个人“值多少钱”的主要尺度?** 现代工资社会不仅通过工资分配生产和消费,还通过市场价格把人的能力、职业、外貌、时间、照护能力和其他可交换特征转化为不同形式的“价值”。这种定价并非全部由资本直接创造,却会进入职业分工、家庭分工与亲密关系。当这些经济约束与传统性别角色结合时,男女可能开始以不同方式承担并交换“可定价价值”:一方更强调收入、资产与资源供给,另一方更容易被要求承担生育、照护、外貌或家庭劳动等成本。本文不把这一结构性分析当作已经被本论文充分证明的经验事实,而把它作为无人化之后值得研究的制度推论:如果社会逐渐削弱人的生产价值对生存资格的支配,那么性别关系中围绕资源、收入与生存保障形成的部分交换结构也可能随之弱化。 由此,本文将“人的价值重构”纳入总论。这里所说的重构并不是宣称一个人“毫无贡献也自动拥有无限价值”,而是提出一个更基础的规范问题:**人的生存资格是否应该依赖其在资本市场上的可定价生产效率?** 当无人化使大量人类劳动变得不再稀缺时,社会可能第一次大规模面对所谓“无用阶层”问题——这里的“无用”不是人的实际价值,而是指一个人对资本增殖而言不再具有足够高的市场需求。届时,社会不得不回答:如果资本不给你定价了,一个人的价值究竟来自哪里? 第四层是“生活”。如果生存逐渐脱离工资,人类需要的不只是公共供给,还需要新的生活关系。本文因此保留AI关系经济学,但明确将其定位为未来生活蓝图而非无人化因果链的必然一环。持续人工主体可以拥有身份、时间、记忆、关系历史以及行动/不行动能力,并通过脑身解耦和多身体具身进入个人、家庭、汽车、家电与其他终端。由此,AI可能从一次性工具逐渐成为长期关系主体,并在未来的两人关系中形成一种可能的“人+人+AI”结构:AI可以承担共同记忆、表达辅助、冲突调和、关系预演和互动训练等功能,但其权力必须被退出权、迁移权、数据权和人格连续性的制度安排所约束。 贯穿这些阶段的深层概念是“解耦”:生产与大规模人类劳动的解耦,基本生活与工资的解耦,部分家庭与生存交换的解耦,以及人工主体与单一身体的解耦。它们共同指向一个更深的历史可能性:**技术进步也许不仅是在减少人必须做多少工作,而是在逐渐减少一个人必须靠什么东西来证明自己“有资格生活”和“值得被爱”。** 本文不预言某一种制度必然到来,也不认为资本主义一定在某个具体年份终结。本文提出的是一套需要未来十年、数十年甚至更长周期检验的历史研究纲领。制度是否发生实质变化,最终不取决于它被称为社会主义、共产主义、福利资本主义还是其他名称,而取决于它能否在新的生产条件下持续组织生产、分配与社会再生产,并把生产力转化为多数人的安全、自由时间、尊严与选择权。 最终,本文试图回应的“未来焦虑”不是“机器会不会取代人”,而是:**如果机器真的让人类不再需要像过去那样大量劳动,人类为什么必须因此失去价值?** 也许无人化革命最深刻的意义不是把人从生产中删掉,而是让人类第一次拥有足够的物质条件,去把“人的价值”与“资本对人的价格”分开。 ## English Artificial intelligence, robotics, intelligent agents, autonomous infrastructure, and unmanned production are jointly creating a productive force that differs from earlier waves of technological upgrading. Previous industrial revolutions largely expanded human mechanical power, energy use, computation, and communication. The deeper potential change of autonomous production is that increasing amounts of production may be completed without proportional increases in paid human labour. The question is therefore not merely whether AI will take jobs, but whether, once large-scale human labour becomes less necessary for production, labour must still remain the principal intermediary through which social members obtain access to social wealth. This paper builds a broader historical framework on the three preceding papers. The first layer is deconstruction. Competitive automation is often privately rational, yet generalized automation may create structural tension among declining labour income, concentrated automation gains, and aggregate demand. If automation gains accrue disproportionately to owners with lower propensities to consume, while falling prices, new tasks, investment, exports, fiscal transfers, and other sources of purchasing power are insufficient to compensate for falling labour income, privately chosen automation may exceed the social optimum. The paper treats the “demand externality of competitive automation” as a conditional and falsifiable mechanism rather than a law of inevitable capitalist collapse. The second layer is construction. The same autonomous productive force may alter the cost structure of education, health care, care, administration, transport, and other essential services. Traditional service providers must themselves live, so housing, family obligations, spousal employment, children’s schooling, and locational preferences become embedded in service costs and geography. Autonomous production does not make services costless; it may shift them from being attached to people who must live at the site of provision toward deployable infrastructure. Some necessities may therefore move from “purchased after earning wages” toward “accessible by virtue of social membership,” creating the possibility of access-based socialisation. The third layer raises a question deeper than institutional naming: **if a person’s survival increasingly no longer depends on the amount of labour they can sell, can capital continue to function as the principal scale for determining what that person is “worth”?** The wage society not only organizes production and consumption through wages; it also translates human capabilities, occupations, appearance, time, care work, and other exchangeable characteristics into different forms of market value. Such pricing is not created by capital alone, but it enters occupational and family divisions as well as intimate relationships. Where these economic constraints interact with inherited gender roles, men and women may bear and exchange “priced value” in different ways: income, assets, and resource provision on one side; reproductive, care, appearance, or household labour on the other. This paper does not present that structural interpretation as an empirically established result of the present study. It treats it as an institutional implication worth testing if autonomous production progressively weakens the power of productive market value over subsistence. The paper therefore introduces the reconstruction of human value as a fourth analytical question. This does not mean claiming that a person has “infinite value” regardless of contribution. It asks a more basic normative question: **should a person’s right to live depend on their market-priced productive efficiency?** As autonomous production makes human labour less scarce, society may confront a large “useless class”—not useless in any human sense, but increasingly “useless” to capital accumulation because the market no longer has sufficient demand for their productive labour. Society would then have to answer: if capital no longer prices you highly, where does human worth come from? The fourth layer is life. If subsistence becomes less dependent on wages, humans need not only public provision but new forms of life and relationship. AI Relational Economics is therefore retained, but explicitly positioned as a possible blueprint for future life rather than a necessary causal step of autonomous production. Persistent artificial subjects may possess identity, time, memory, relational history, and the capacity for action or non-action, entering personal, family, automotive, appliance, and other everyday environments through brain–body decoupling and multi-embodiment. They may eventually participate in a possible “human + human + AI” relational structure, supporting shared memory, communication, conflict mediation, relationship rehearsal, and interaction training. Such systems, however, would require strong rights of exit, migration, portability, and control over relational data and continuity. A deeper concept connects these stages: decoupling—the decoupling of production from large-scale human labour, of basic life from wages, of parts of family relations from subsistence exchange, and of artificial subjects from single bodies. Together these point toward a deeper historical possibility: **technological progress may not only reduce how much humans have to work, but gradually reduce the things a person must use to prove that they deserve to live and deserve to be loved.** The paper does not predict one inevitable institution, nor does it predict that capitalism will end in any particular year. It proposes a research programme that may require a decade, several decades, or longer to evaluate. Whether institutions change in substance will ultimately depend less on whether they are called socialism, communism, welfare capitalism, or something else than on whether they can continue to organize production, distribution, and social reproduction under new productive conditions and convert productive power into security, free time, dignity, and freedom of choice for the majority. The deepest question of “future anxiety” is therefore not simply whether machines will replace humans, but: **if machines really allow humanity to work far less than before, why must that future make humans less valuable?** The autonomous revolution may ultimately matter not because it removes humans from production, but because it may provide the material conditions under which humanity can finally separate “human worth” from “capital’s price on a person.”
No takes yet. Share an insight, caveat, or question.
哲宏 刘 (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: