AI关系经济学:从持续人工主体到多身体智能生态
Abstract
中文摘要:本文综述并进一步扩展一条正在形成的产品与产业路线:从持续人工主体(Persistent Artificial Subject)出发,通过脑身解耦(Brain–Body Decoupling)与多身体具身(Multi-Embodiment),将 AI 从一次性任务工具逐步转化为可以长期存在于个人、家庭、汽车、家电和其他生活终端中的人工关系主体,并进一步提出"AI关系经济学"(AI Relational Economics)作为描述这一潜在经济结构的概念框架。此前的伴侣机器人范式提出:如果产品的核心目标是陪伴,就没有必要首先解决开放世界通用机器人问题。伴侣机器人可以成为一种以陪伴为核心的专用具身产品;其"脑"由持续运行的人工主体承担,包括身份、人格、记忆、关系历史、时间感知和环境感知,而身体主要承担感知、表达和执行。Project Sue 以数字环境为参考,将 Identity、Perception、Self-Interpretation、Intention、Action/Non-Action、Experience、Memory、Trait Dynamics 和 Updated Identity 组织成持续主体闭环。本文进一步扩展三个方向。第一,从单一用户群扩展到人生连续性。 对有社交意愿的年轻人,伴侣机器人可以是"家里的伙伴 + 未来真人关系的人际预训练";对不希望扩大社交的人,可以主要提供陪伴与安抚;对具有较高生活照护需求的独居人士,则进一步提供陪伴、安抚与照护。更重要的是,同一个人工主体可以跨越十年以上的人生阶段,随着用户年龄、家庭结构和社交目标变化而改变价值,而不必重新建立一段关系。第二,从个人伴侣扩展到家庭主体。 用户建立家庭后,原本属于个人的人工主体可以自然进入传统家庭:在家中通过机器人身体提供家务与陪伴;一家人出行时,则可以通过汽车的车机、麦克风、摄像头和其他接口继续存在。主体保持连续,身体发生切换。第三,从产品经济扩展到关系经济。 如果人格、记忆和关系历史构成用户所认识的"她"的连续性,那么长期购买的就不再只是某一个模型、机器人或设备,而可能是一种持续运行的关系服务。身体可以升级,大模型可以更换,终端可以迁移,但主体的身份、记忆和关系历史可以保持连续。由此产生"关系连续性""关系资本""具身迁移"等概念,并形成大模型公司、机器人公司、家电公司、汽车公司及其他终端企业共同参与的多身体 AI 生态。本文因此提出一个更大的商业与产业假设:伴侣机器人可能不是通用机器人的终点,而是持续人工主体进入 C 端生活的入口;家庭伙伴可能是个人伴侣自然扩展后的下一阶段;多身体具身可能是一种新的产业架构;而 AI关系经济学,则试图描述以持续人工主体和长期关系价值为核心的新型经济活动。上述观点均属于需要通过原型、用户研究、商业实验、长期部署和产业协作验证的产品与经济假设。英文 Abstract:This paper reviews and extends a product and industrial trajectory: starting from the Persistent Artificial Subject, using Brain–Body Decoupling and multi-embodiment to transform AI from a task-oriented tool into a long-running artificial relational subject that can persist across personal devices, family robots, cars, appliances, and other everyday interfaces, and proposing "AI Relational Economics" as a framework for the possible economic structure around such subjects.The earlier companion-robot paradigm proposed that if companionship is the primary product function, open-world general-purpose robotics does not have to be solved first. A companion robot can instead be a purpose-specific embodied product centered on companionship. Its "brain" is a persistent artificial subject containing identity, personality, memory, relational history, temporal awareness, and environmental awareness, while the body provides perception, expression, and execution. Project Sue serves as a digital reference for organizing Identity, Perception, Self-Interpretation, Intention, Action/Non-Action, Experience, Memory, Trait Dynamics, and Updated Identity into a persistent-subject loop.This paper extends the framework in three directions: life-stage continuity; expansion from personal companion to family subject; and the transition from product economics to relational economics.The central hypothesis is:Companion robotics may be an entry point for persistent artificial subjects into consumer life; family partnership may be their natural next stage; multi-embodiment may become a new industrial architecture; and AI Relational Economics may describe an emerging economy centered on persistent artificial subjects and long-term relational value.These remain product and economic hypotheses requiring validation.
Fields of Study & Subjects
Embodied and Extended Cognition
Economic and Technological Innovation
Innovation
Sustainability
Human-Machine Systems
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 24 August 2026
- Language
- CMN
- Document type
- Preprint