基于AIGC的电商智能文案对消费者购买意愿的影响：在整合性理论分析框架视角下的研究

The Impact of Generative Artificial Intelligence (AIGC)-Based Smart Copywriting on Consumer Purchase Intention in E-Commerce: A Study from the Perspective of an Integrated Theoretical Analysis Framework

摘要: 生成式人工智能技术的爆发式应用正在重塑电商内容生产范式，但其缺乏人类主体意识的生产黑箱特性使得基于传统人际信任的在线营销理论面临解释困境。本文立足于相关管理办法实施的制度背景，整合信息系统、消费者行为学与技术伦理视角，构建了一个整合性理论分析框架，旨在揭示智能文案对消费者购买意愿的深层影响机制。研究表明，智能文案作为外部刺激，通过激发消费者感知有用性、多维信任与感知风险的动态博弈共同作用于购买意愿。其中，生成式人工智能的高信息质量是建立认知信任的基石，但其情感模拟的有效性受制于人工智能可能存在的情感认知障碍；来源披露在提升算法透明度与决策自主性的同时，可能因破坏沉浸感而触发恐怖谷效应，进而加剧感知风险。进一步的理论推演发现，产品类型差异与平台责任归因显著调节上述路径，实用品消费更依赖于基于信息的内化机制，而享乐品消费则更敏感于基于情感的认同缺失。本研究不仅揭示了人机协同语境下消费决策中信任构建与风险感知的非零和关系，也为电商平台建立分级披露机制以及商家制定差异化的人机协同内容策略提供了理论依据与实践启示。

Abstract: The explosive application of generative artificial intelligence is reshaping the e-commerce content production paradigm. However, its black-box nature, lacking human consciousness, challenges traditional online marketing theories based on interpersonal trust. Grounded in the institutional context of relevant regulatory measures, this paper integrates perspectives from information systems, consumer behavior, and technology ethics to construct an integrated theoretical framework revealing the underlying mechanisms through which smart copywriting affects consumer purchase intentions. Research indicates that smart copywriting acts as an external stimulus, shaping purchase intentions through a dynamic interplay of consumer perceived usefulness, multidimensional trust, and perceived risk. The high information quality of generative artificial intelligence forms the cornerstone of cognitive trust, yet its emotional simulation effectiveness is constrained by potential emotional cognitive barriers within the artificial intelligence. Source disclosure enhances algorithmic transparency and decision-making autonomy but may simultaneously disrupt immersion and trigger the uncanny valley effect, thereby exacerbating perceived risks. Further theoretical deduction reveals that product types and platform responsibility attribution significantly moderate these pathways. Utilitarian consumption relies heavily on information-based internalization mechanisms, whereas hedonic consumption is highly sensitive to the absence of emotional identification. This study uncovers the non-zero-sum relationship between trust building and risk perception in consumer decision-making under a human-machine collaboration context. It provides theoretical foundations and practical insights for e-commerce platforms to establish tiered disclosure mechanisms and for merchants to develop differentiated human-machine collaborative content strategies.