arXiv cs.CY· Leonardo Borges (Eigenstack Pty Ltd), Asif Q. Gill (University of Technology Sydney)·· 2 天前AI 评分24
从工具到社交实体：共享社交空间中个性化 AI 智能体的准社会动态
The Tool-to-Entity Threshold: Parasocial Dynamics of Personalised AI Agents in Shared Social Spaces
AI 导读
研究提出"身份标记框架"，用命名、视觉身份、通讯录存在、人格来源、社交共现与持久性六个设计变量，解释 AI 智能体何时从用户操作的工具被重新归类为可建立关系的社交实体，且该转变独立于模型能力。
正文 · 原文
Abstract:As AI agents acquire names, avatars, phone numbers, and persistent personalities, they increasingly inhabit the same messaging platforms and group conversations as the humans they serve, crossing from tools their users operate into social entities their users relate to. Yet no existing framework identifies the infrastructural markers that cause the shift: the anthropomorphism literature catalogues perceptual cues rendered inside the interaction surface, not the agent's placement in the user's social and computational graph. We propose the identity marker framework: six design variables - naming, visual identity, contact presence, personality derivation, social co-presence, and persistence - that collectively trigger a psychological reclassification, categorical in its consequences even where thetransition itself is gradual, and operating independently of model capability. We read the framework through parasocial interaction theory (Horton & Wohl, 1956) and the Computers Are Social Actors paradigm (Nass et al., 1994), and identify four novel dynamics that arise when such an agent joins existing group conversations: bidirectional information asymmetry, delegation legibility, social norm negotiation, and parasocial contagion. Our method is autoethnographic: the first author built and deployed a personalised agent into WhatsApp and Signal group chats over two months of live use, supplemented by twelve structured interviews with the group members who encountered it. We treat this as a preliminary qualitative evaluation of the framework, with controlled experimental validation set out as future work. The strongest design implication runs through all four dynamics: in shared social spaces, consent to an agent's presence is categorically distinct from consent to its processing of the messages exchanged there, and existing consent frameworks collapse the two.
| Comments: | 25 pages, 3 figures, 2 tables. v3: ChatGPT memory re-characterised as dyadic and user-owned rather than stateless (§1, Table 2, §7.1); adds OpenAI (2025) citation. v2: corpus message count corrected to ~16,000; threshold claim reframed as categorical rather than discrete; adds §5.4 and H4. Scope and consent basis unchanged |
| Subjects: | Human-Computer Interaction (cs.HC); Computers and Society (cs.CY) |
| Cite as: | arXiv:2608.13586 [cs.HC] |
| (or arXiv:2608.13586v3 [cs.HC] for this version) | |
| https://doi.org/10.48550/arXiv.2608.13586 arXiv-issued DOI via DataCite |
Submission history
From: Leonardo Borges Barbosa [view email]
[v1]
Fri, 10 Jul 2026 12:43:52 UTC (900 KB)
[v2]
Wed, 19 Aug 2026 02:30:04 UTC (540 KB)
[v3]
Wed, 14 Oct 2026 15:26:50 UTC (926 KB)
来源：arXiv cs.CY · arxiv.org