中国健康心理学杂志

目的:人工智能(AI)技术深度融入数字传播生态的背景下，身体符号的技术再编码过程日益显著，肥胖形象逐步成为短视频平台的视觉焦点之一。本研究揭示AI驱动的“变胖”功能如何促成肥胖污名的视觉化、符号化传播，并探讨污名个体在新媒体语境中的应对策略，从而扩展污名管理沟通模型(SMC)在AI视觉传播环境中的理论适用性与解释力。方法:采用质性研究为主，运用深度访谈、数字民族志等方法，围绕2023年10月至2024年1月抖音等平台上关于“AI变胖”内容开展数据采集与分析。研究对象涵盖60个短视频用户账号及48位具有肥胖标签的受访者，借助污名沟通模型(MSC)理论中“标记-责任-社会危险”3个维度展开分析。结果:AI技术在新媒体平台中通过“无头化”处理、身体夸张变形与审美算法强化等手段，构建了“视觉污名的结构性算法机制”。该机制不仅加剧了对肥胖者的物化呈现与审美排斥，也助长了“胖即病”“胖即失败”的社会刻板印象。面对污名，肥胖个体主要展现出3种沟通策略：消极回避、才艺补偿与自我赋权；其中，获得社交支持与主动话语反击者，其社会认同与心理恢复效能更显积极。结论:AI视觉技术在污名再生产中已不再是中性工具，而是构成污名传播的结构性变量之一。新媒体平台的算法偏好、语境操控与话语权分配机制对污名的形成与扩散具有深远影响。本研究在丰富SMC理论本土化应用的同时，呼吁平台强化伦理治理责任，通过算法优化、内容干预与审美多元引导共同遏制身体污名的媒介扩散，为构建包容性数字传播环境提供理论支持与实证依据。

Objective:In the context of artificial intelligence(AI)technologies becoming deeply embedded in the digital communication ecology, the symbolic reconfiguration of the human body has gained increasing prominence, with obesity imagery emerging as a focal point in algorithmically curated short video content.This study investigates how AI-enabled "fat transformation" functions contribute to the visual and symbolic reproduction of obesity stigma, and how such technologically mediated portrayals reinforce aesthetic hierarchies and discriminatory social perceptions.Methods:Drawing on the Stigma Management Communication(SMC)framework, the study employed a qualitative methodology-primarily in-depth interviews and digital ethnography-to analyze data collected from 60 short video user accounts and 48 individuals experiencing obesity-related stigma on platforms such as Douyin between October 2023 and January 2024.The analysis was guided by the "marking, " "responsibility, " and "social danger" dimensions of the Message-Based Stigma Communication(MSC)model.Results:The findings revealed that AI visual technologies constructed a structural algorithmic mechanism of stigma reproduction, characterized by facial occlusion, exaggerated bodily distortion, and algorithmic reinforcement of dominant aesthetic norms.These mechanisms contributed to the objectification and marginalization of obese individuals, while simultaneously normalizing stigmatizing narratives such as "fat equals failure" or "fat equals disease." In response, stigmatized individuals adopted coping strategies such as passive avoidance, talent-based compensation, and self-empowerment, with social support and counter-discursive expression proving especially effective in mitigating psychological harm.Conclusion:The study concludes that AI-driven media technologies are not neutral tools but constitute structural variables in the propagation of stigma.It further argues that platform-level algorithmic governance, discursive regulation, and ethical design interventions are necessary to counteract the mediatized amplification of body-related stigma and to foster a more inclusive digital communication environment.By localizing and extending the theoretical scope of the SMC framework, this research offers both conceptual innovation and practical implications for digital media governance in the age of intelligent visual technologies.