工具理性与情感质疑:AI伴侣(AIC)用户态度的二元分化分析
Instrumental Rationality and Emotional Skepticism: A Binary Analysis of User Attitudes Toward AI Partners (AIC)
AI伴侣(AI Companion,AIC)作为生成式AI落地情感陪伴的典型形态,已成为数字社会中缓解孤独、提供情绪支持的重要载体。本文将以工具理性与情感质疑的二元分化命题为核心,以梳理中英文核心文献为依据,阐述AIC用户态度二元分化的核心表征、形成机理与现实影响。分析发现公众认可AIC的工具价值,却质疑其情感真实与关系双向;AIC即提升工具使用意愿,制约深度情感的建立,同时也带来社交现实关系的弱化分险。最后,本文将从技术透明、认知引导、伦理规范三个维度提出优化路径,为AIC健康发展与理性人际关系领域的完善提供参考。
As a paradigm of generative AI in emotional companionship, AI Companion (AIC) has become a vital digital solution for alleviating loneliness and providing emotional support. This study examines the dualistic dichotomy between instrumental rationality and emotional skepticism through a systematic review of core Chinese and English literature, elucidating the core manifestations, formation mechanisms, and practical impacts of AIC user attitudes. The analysis reveals that while the public recognizes AIC’s instrumental value, they question its emotional authenticity and bidirectional relationship-building capabilities. AIC enhances tool usage willingness but constrains deep emotional connections, while simultaneously weakening the risk-sharing mechanisms of real-world social relationships. Finally, the paper proposes optimization pathways from three dimensions—technical transparency, cognitive guidance, and ethical norms—to guide the healthy development of AIC and the refinement of rational interpersonal relationships.
| [1] |
腾讯研究院. 十问“AI陪伴”研究报告:现状、趋势与机会. 2024(10). |
| [2] |
黄浩宇, 任梓源, 方兴东. 算法的逻辑与人类的温度—2025年情感智能研究综述[J/OL]. 青年记者. https://doi.org/10.15997/j.cnki.qnjz.20260107.001 |
| [3] |
赵瑜, 曹凌霄. 人机亲密关系中的“媒介唤起”与用户试探行为研究[J]. 新闻与传播研究, 2026(1):78-95. |
| [4] |
张丽萍, 张立涛. 对“生成式情绪”的共情实践:人对AI情感智能体的情绪感知与共情[J]. 国际新闻界, 2025(7). |
| [5] |
刘梦雪, 冯雨奂. AIGC时代青年与智能伴侣的虚拟交互及其风险审视[J]. 新疆社会科学, 2024(6). |
| [6] |
工业和信息化部等部门. 人工智能科技伦理管理服务办法(试行)[EB/OL].(2025-08-22). https://www.miit.gov.cn/jgsj/kjs/jscx/gjsfz/art/2025/art_092a447008f340d3abd55819b8c8e5cf.html. |
| [7] |
Joi AI. Z世代AI伴侣情感关系调查报告[R]. 2025. |
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
/
| 〈 |
|
〉 |