联系我们

论文查询结果	返回搜索

论文编号：	15400
作者编号：	2120233656
上传时间：	2025/6/10 11:57:25
中文题目：	社交电商平台用户 AIGC 信息规避行为研究
英文题目：	Research on Users’ Avoidance Behavior Toward AI-Generated Content in Social Commerce Platforms
指导老师：	石鉴
中文关键字：	人工智能生成内容；信息规避行为；社交电商；认知-情感-意动框架
英文关键字：	Artificial Intelligence Generated Content? Information Avoidance Behavior? Social Commerce? Cognitive-Affective-Conative Framework
中文摘要：	随着生成式人工智能技术的快速发展，人工智能生成内容（AIGC）正通过 文本生成、图像合成等多元化形式重构内容创作范式。现有研究表明，社交电 商平台已成为 AIGC 技术落地的重要应用场景，其在提升内容生产效率与个性 化推荐能力的同时，也衍生出信息真实性危机、用户信任机制弱化及技术伦理 争议等新型治理挑战。本研究以“社交电商平台用户 AIGC 信息规避行为”为核 心议题，综合运用混合研究方法，系统探讨用户对 AIGC 信息的认知评价、情感 反应与行为决策的关联机制，旨在揭示其规避行为的形成路径与影响因素。 在质性研究阶段，本研究通过对 18 名 AI 应用高频用户进行半结构化访 谈，采用程序化扎根理论三级编码，提炼出 AIGC 感知过载、AIGC 感知可控度、 AIGC 感知有用性、AI 焦虑、AI 疲劳及 AIGC 信息规避行为 6 个核心范畴，并构 建形成机制理论模型。量化研究阶段，整合情感认知理论、技术接受模型、不确 定性管理理论及认知-情感-意动（CAC）框架，设计涵盖 6 个潜变量、11 条作用 路径的结构方程模型，通过 612 份有效问卷数据验证假设。实证结果表明：（1） AIGC 感知过载对 AI 焦虑（β=0.259）、AI 疲劳（β=0.447）及 AIGC 信息规避行 为（β=0.321）均呈显著正向影响；（2）AIGC 感知可控度对 AI 焦虑（β=-0.221） 与 AI 疲劳（β=-0.254）具有显著负向调节作用；（3）AI 焦虑与 AI 疲劳不仅直接 驱动 AIGC 信息规避行为（β=0.243；β=0.255），还形成情感闭环反馈机制，强 化规避倾向。 本研究从消极行为视角切入，系统揭示了社交电商场景下用户 AIGC 信息 规避行为的形成机理。理论层面，区别于传统信息规避行为研究对 UGC 与 PGC 的路径依赖，研究结合 AIGC 技术特征与社交电商场景独特性构建了基于 CAC 框架的动态模型，验证了 CAC 理论在 AIGC 信息行为研究中的适用性，为理解 用户技术接受与行为决策的交互机制提供了新视角。实践层面，为平台提出分 级信息过滤、AIGC 内容标注及用户教育策略，助力平衡技术效率与伦理风险。 研究成果可为平台平衡技术效率与内容生态治理提供参考，同时为 AIGC 技术 的规范化应用与用户行为研究提供实证支持。
英文摘要：	With the rapid advancement of generative artificial intelligence (AI) technologies, AI-generated content (AIGC) is reshaping content creation paradigms through diverse forms such as text generation and image synthesis. Existing studies indicate that social e-commerce platforms have emerged as a crucial application scenario for AIGC technologies. While enhancing content production efficiency and personalized recommendation capabilities, these technologies have also given rise to governance challenges such as information authenticity crises, weakened user trust mechanisms, and ethical controversies surrounding technology. This study centers on the ”AIGC information avoidance behavior of social e-commerce platform users.” Through a mixed-methods research approach, it systematically explores the associative mechanisms among users’ cognitive evaluation, emotional responses, and behavioral decisions toward AIGC information, aiming to reveal the formation pathways and influencing factors of avoidance behavior. In the qualitative research phase, semi-structured interviews were conducted with 18 AI application high-frequency users. Using the three-level coding approach of grounded theory, six core categories were identified: AIGC perception overload, perceived controllability, perceived usefulness, AI anxiety, AI fatigue, and AIGC information avoidance behavior. A theoretical model of the formation mechanism was subsequently constructed. In the quantitative research phase, the study integrates the affective-cognitive theory, technology acceptance model (TAM), uncertainty management theory, and the cognitive-affective-conative (CAC) framework to design a structural equation model comprising six latent variables and eleven causal pathways. The hypotheses were validated using data from 612 valid questionnaires. The empirical results show that: (1) AIGC perception overload has a significant positive impact on AI anxiety (β=0.259), AI fatigue (β=0.447), and AIGC information avoidance behavior (β=0.321)? (2) Perceived controllability of AIGC significantly negatively moderates AI anxiety (β=-0.221) and AI fatigue (β=-0.254)? (3) AI anxiety and AI fatigue not only directly drive AIGC in formation avoidance behavior (β=0.243? β=0.255) but also form an emotional feedback loop that reinforces avoidance tendencies. From the perspective of negative behavior, this study systematically reveals the formation mechanism of AIGC information avoidance behavior in the context of social e-commerce. Theoretically, it differentiates itself from traditional information avoidance research that is path-dependent on UGC and PGC by incorporating the unique characteristics of AIGC technology and the specificity of social e-commerce scenarios. The study constructs a dynamic model based on the CAC framework, validating the applicability of CAC theory in AIGC information behavior research and offering a new perspective for understanding the interactive mechanisms of user technology acceptance and behavioral decision-making. Practically, the study proposes strategies for hierarchical information filtering, AIGC content labeling, and user education to help platforms balance technological efficiency and ethical risks. The findings provide valuable insights for platforms to manage the balance between technological efficiency and content ecosystem governance, as well as empirical support for the standardized application of AIGC technology and research on user behavior.formation avoidance behavior (β=0.243? β=0.255) but also form an emotional feedback loop that reinforces avoidance tendencies. From the perspective of negative behavior, this study systematically reveals the formation mechanism of AIGC information avoidance behavior in the context of social e-commerce. Theoretically, it differentiates itself from traditional information avoidance research that is path-dependent on UGC and PGC by incorporating the unique characteristics of AIGC technology and the specificity of social e-commerce scenarios. The study constructs a dynamic model based on the CAC framework, validating the applicability of CAC theory in AIGC information behavior research and offering a new perspective for understanding the interactive mechanisms of user technology acceptance and behavioral decision-making. Practically, the study proposes strategies for hierarchical information filtering, AIGC content labeling, and user education to help platforms balance technological efficiency and ethical risks. The findings provide valuable insights for platforms to manage the balance between technological efficiency and content ecosystem governance, as well as empirical support for the standardized application of AIGC technology and research on user behavior.
查看全文：	预览  下载(下载需要进行登录)