人工智能生成评论总结如何影响购买意愿？——来自眼动实验的证据

摘要/Abstract

摘要： 在信息过载的数字消费时代,人工智能生成评论总结(AIGRS)已成为电商平台提高消费者决策效率的核心工具,但其视觉呈现策略对购买意愿的影响机制仍处于理论“黑箱”状态。本研究基于精细加工可能性模型、顺序效应与视觉注意力理论,构建了AIGRS效价顺序(正—负vs.负—正)与呈现结构(列表体vs.叙事体)的交互框架,并引入产品类型(搜索品vs.体验品)作为调节变量,通过眼动追踪实验探究了AIGRS影响购买意愿的内在机制。研究发现:(1)AIGRS“效价顺序—呈现结构—产品类型”对正面内容视觉注意的影响具有匹配效应,即搜索品适配“列表体与正—负效价”组合,体验品适配“叙事体与负—正效价”组合;(2)消费者对正面内容的视觉注意在上述交互效应与购买意愿之间发挥完全中介作用。研究结论不仅从生理层面揭示了“信息呈现—注意力分配—决策行为”的传导机制,为AIGC营销、信息传播与消费者决策交叉领域贡献了实证证据与理论增量,更为电商平台优化算法设计与提升购买转化率提供了具有场景适配性的差异化实践路径。

关键词: 人工智能, AIGC, 在线评论, 眼动实验, 购买意愿

Abstract: In the digital consumption era characterized by information overload, AI-generated review summaries (AIGRS) have become a core tool for e-commerce platforms to improve consumers' decision-making efficiency, yet the impact of their visual presentation on purchase intentions remains a theoretical “black box”. Based on the elaboration likelihood model, order effects, and visual attention theory, this study examines the interactive effects of AIGRS valence sequence (positive-negative vs. negative-positive) and presentation structure (list-based vs. narrative-based), with product type (search goods vs. experience goods) as a moderator. An eye-tracking experiment reveals a matching effect on visual attention to positive content: search goods are better aligned with a list-based, positive-negative sequence, whereas experience goods are better aligned with a narrative-based, negative-positive sequence. Moreover, visual attention to positive content fully mediates the interactive effects on purchase intention. These findings elucidate the attention-based mechanism linking information presentation to consumer decision-making and offer actionable insights for optimizing AIGRS design in e-commerce platforms.

Key words: Artificial Intelligence, AIGC, Online Reviews, Eye-Tracking Experiment, Purchase Intention

中图分类号:

F713.55

张鹏, 刘春慧. 人工智能生成评论总结如何影响购买意愿？——来自眼动实验的证据[J]. 财经论丛, 2026, 42(6): 101-112.

ZHANG Peng, LIU Chunhui. How Do AI-Generated Review Summaries Influence Purchase Intention: Evidence from an Eye-Tracking Experiment[J]. Collected Essays on Finance and Economics, 2026, 42(6): 101-112.

参考文献

[1] 高海霞, 雷超龙. 负面评论的反直觉效应: 研究述评与展望[J]. 外国经济与管理, 2025, (2): 103-118.
[2] Jia S. J., Chi O. H., Chi C. G. Unpacking the Impact of AI vs. Human-Generated Review Summary on Hotel Booking Intentions[J]. International Journal of Hospitality Management, 2025, 126: 104030.
[3] Bigne E., Chatzipanagiotou K., Ruiz C. Pictorial Content, Sequence of Conflicting Online Reviews and Consumer Decision-Making: The Stimulus-Organism-Response Model Revisited[J]. Journal of Business Research, 2020, 115: 403-416.
[4] Wang Y., Zhong K., Liu Q. Let Criticism Take Precedence: Effect of Side Order on Consumer Attitudes Toward a Two-Sided Online Review[J]. Journal of Business Research, 2022, 140: 403-419.
[5] Huang G. H., Chang C. T., Bilgihan A., et al. Helpful or Harmful？A Double-Edged Sword of Emoticons in Online Review Helpfulness[J]. Tourism Management, 2020, 81: 104135.
[6] Hamby A., McFerran B., Fuller C. The Power of Proximity: Exploring Narrative Language in Consumer Reviews[J]. Journal of Marketing, 2026, 90(2): 157-176.
[7] Nelson P. Advertising as Information[J]. Journal of Political Economy, 1974, 82(4): 729-754.
[8] 郭婷婷, 李宝库. “看得见”还是“摸得着”？——在线评论中感官线索引发的意象体验效应[J]. 财经论丛, 2019, (9): 82-91.
[9] Foo L. G., Rahmani H., Liu J. AI-Generated Content (AIGC) for Various Data Modalities: A Survey[J]. ACM Computing Surveys, 2025, 57(9): 1-66.
[10] El-Kassas W. S., Salama C. R., Rafea A. A., et al. Automatic Text Summarization: A Comprehensive Survey[J]. Expert Systems with Applications, 2021, 165: 113679.
[11] Na J., Long R., Chen H., et al. Sentiment Analysis of Online Reviews of Energy-Saving Products Based on Transfer Learning and LBBA Model[J]. Journal of Environmental Management, 2024, 360: 121083.
[12] Seo H. J., Song T. H. Do All the Service Attributes Matter？Application of Customer Review-Based Attribute Extraction Considering Brand Status[J]. Journal of Retailing and Consumer Services, 2025, 87: 104345.
[13] Shi Y., Shang M., Qi Z. Intelligent Layout Generation Based on Deep Generative Models: A Comprehensive Survey[J]. Information Fusion, 2023, 100: 101940.
[14] Petty R. E., Cacioppo J. T. The Elaboration Likelihood Model of Persuasion[J]. Advances in Experimental Social Psychology, 1986, 19: 123-205.
[15] 鄢慧丽, 廖其威, 熊浩. 情感还是非情感？表情符号对餐厅在线评论有用性的影响[J]. 旅游学刊, 2024, (5): 145-160.
[16] Petty R. E., Cacioppo J. T., Schumann D. Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement[J]. Journal of Consumer Research, 1983, 10(2): 135-146.
[17] Hamby A., Daniloski K., Brinberg D. How Consumer Reviews Persuade Through Narratives[J]. Journal of Business Research, 2015, 68(6): 1242-1250.
[18] Haugtvedt C. P., Wegener D. T. Message Order Effects in Persuasion: An Attitude Strength Perspective[J]. Journal of Consumer Research, 1994, 21(1): 205-218.
[19] Zuschke N. Order in Multi-Attribute Product Choice Decisions: Evidence from Discrete Choice Experiments Combined with Eye Tracking[J]. Journal of Behavioral Decision Making, 2023, 36(4): e2320.
[20] Petty R. E., Tormala Z. L., Hawkins C., et al. Motivation to Think and Order Effects in Persuasion: The Moderating Role of Chunking[J]. Personality and Social Psychology Bulletin, 2001, 27(3): 332-344.
[21] Lu S., Qiu L., Wang K. The Effects of the Format of Two-Sided Online Reviews: A Linguistic Perspective[J]. Information & Management, 2021, 58(8): 103554.
[22] 张秋杭. 句法从属与语篇设景机制——以汉语描写性关系从句为例[J]. 解放军外国语学院学报, 2023, (2): 77-84.
[23] 张健军. 转折复句的定景机制及其研究意义[J]. 语言教学与研究, 2014, (2): 68-77.
[24] 张湘一, 吴一琳. 视觉注意对决策的影响及其作用机制[J]. 心理科学进展, 2024, (11): 1829-1843.
[25] Chen L., Jing K., Mei Y. The Effect of Consumption Goals on Review Helpfulness: Behavioral and Eye-Tracking Research[J]. Journal of Retailing and Consumer Services, 2024, 76: 103563.
[26] Yang S. F. An Eye-Tracking Study of the Elaboration Likelihood Model in Online Shopping[J]. Electronic Commerce Research and Applications, 2015, 14(4): 233-240.
[27] Fei M., Tan H., Peng X., et al. Promoting or Attenuating？An Eye-Tracking Study on the Role of Social Cues in E-Commerce Livestreaming[J]. Decision Support Systems, 2021, 142: 113466.
[28] Jin J., Wang A., Wang C., et al. How Do Consumers Perceive and Process Online Overall vs. Individual Text-Based Reviews？Behavioral and Eye-Tracking Evidence[J]. Information & Management, 2023, 60(5): 103795.
[29] Gallo I., Sood S., Mann T. C., et al. The Heart and the Head: On Choosing Experiences Intuitively and Possessions Deliberatively[J]. Journal of Behavioral Decision Making, 2017, 30(3): 754-768.
[30] 韩召, 杜刚, 熊爱华. 电商直播类型对新产品购买意愿的影响[J]. 经济管理, 2023, (4): 133-150.
[31] Naylor R. W., Lamberton C. P., Norton D. A. Seeing Ourselves in Others: Reviewer Ambiguity, Egocentric Anchoring, and Persuasion[J]. Journal of Marketing Research, 2011, 48(3): 617-631.
[32] 龚潇潇, 蒋雪涛, 玉胜贤. AI虚拟主播角色与产品类型对消费者购买意愿的交互影响研究[J]. 管理学报, 2024, (9): 1373-1381.
[33] He S. X., Bond S. D. Why Is the Crowd Divided？Attribution for Dispersion in Online Word of Mouth[J]. Journal of Consumer Research, 2015, 41(6): 1509-1527.
[34] Wang Q., Yang S., Liu M., et al. An Eye-Tracking Study of Website Complexity from Cognitive Load Perspective[J]. Decision Support Systems, 2014, 62: 1-10.
[35] Hayes A. F. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach[M]. New York: Guilford Press, 2013.
[36] Chen Y., Wang H., Hill S. R., et al. Consumer Attitudes Toward AI-Generated Ads: Appeal Types, Self-Efficacy and AI's Social Role[J]. Journal of Business Research, 2024, 185: 114867.
[37] Park J. K., Ahn S. Traditional vs. AI-Generated Brand Personalities: Impact on Brand Preference and Purchase Intention[J]. Journal of Retailing and Consumer Services, 2024, 81: 104009.
[38] Wang Q., Ma D., Chen H., et al. Effects of Background Complexity on Consumer Visual Processing: An Eye-Tracking Study[J]. Journal of Business Research, 2020, 111: 270-280.
[39] 刘洪志, 魏子晗, 盈嘉, 等. 因果与权重: 决策的眼动模型[J]. 心理科学, 2022, (1): 242-249.