人工智能视域下直播电商的研究热点与演化路径
Research Hotspots and Evolution Trajectory of Live-Streaming E-Commerce from an AI Perspective
DOI: 10.12677/ecl.2026.151104, PDF,   
作者: 卜小柯:贵州大学省部共建公共大数据国家重点实验室,贵州 贵阳
关键词: 直播电商BERTopic文本挖掘数字经济乡村振兴Live-Streaming E-Commerce BERTopic Text Mining Digital Economy Rural Revitalization
摘要: 在数字经济已成为高质量发展核心引擎的宏观背景下,直播电商作为其中最具活力的业态,正处于从“流量红利”驱动向“技术红利”驱动的深刻范式转型期。生成式人工智能(AIGC)、大语言模型(LLM)技术的爆发式增长,不仅重塑了内容生产与交互逻辑,更引发了深层的产业变革与伦理挑战。本文在宏观把握数字经济发展脉络的基础上,以CNKI数据库中CSSCI及北大核心期刊收录的1169篇高质量文献为样本,创新性引入中文预训练语言模型BAAI/bge-large-zh进行高维向量化表征,并结合K-Means聚类与BERTopic主题模型开展深度挖掘。研究揭示了该领域形成的五大核心议题群:基础设施层聚焦于智媒体融合传播;平台生态层呈现出向“乡村振兴”与“垂直知识”深耕的趋势;算法机制层探讨推荐算法对用户感知的实证影响;AIGC驱动层突显了从“真人主播”向“虚拟数字人”的演进;伦理治理层则从版权合规延伸至税收征管与跨境数据流动。演化分析表明,研究焦点正由早期的商业模式探讨,向底层的技术伦理反思、人机协同交互及绿色供应链整合纵深发展。
Abstract: Against the macro background in which the digital economy has become a core engine of high-quality development, live-streaming e-commerce—one of the most dynamic business formats—has entered a profound paradigm shift from a “traffic dividend-driven” model to a “technology dividend-driven” model. The explosive growth of generative artificial intelligence (AIGC) and large language models (LLMs) has not only reshaped content production and interaction mechanisms, but also triggered deep industrial transformations and ethical challenges. Grounded in an overarching perspective on the evolution of the digital economy, this study examines 1169 high-quality articles indexed in CSSCI and Peking University Core Journals from the CNKI database. A Chinese pretrained language model (BAAI/bge-large-zh) is introduced to generate high-dimensional semantic embeddings, which are further analyzed using K-Means clustering and the BERTopic topic modeling framework. The results identify five core thematic clusters in this research field: the infrastructure layer focuses on intelligent media integration and communication; the platform ecosystem layer shows a deepening orientation toward rural revitalization and vertical knowledge domains; the algorithmic mechanism layer examines the empirical effects of recommendation algorithms on user perception; the AIGC-driven layer highlights the transition from human anchors to virtual digital humans; and the ethical governance layer extends from copyright compliance to issues of taxation governance and cross-border data flows. Evolutionary analysis indicates that research attention has shifted from early discussions of business models toward in-depth reflections on technological ethics, human-machine collaboration, and the integration of green supply chains.
文章引用:卜小柯. 人工智能视域下直播电商的研究热点与演化路径[J]. 电子商务评论, 2026, 15(1): 851-861. https://doi.org/10.12677/ecl.2026.151104

参考文献

[1] 李佳霖, 温金燕. 数实融合对高质量发展的影响研究[J/OL]. 科研管理: 1-16.
https://link.cnki.net/urlid/11.1567.g3.20250905.1709.004, 2025-12-22.
[2] 张鸿, 王宇翔, 张媛. “十五五”时期国家数字经济发展规划前瞻性研究[J]. 西安财经大学学报, 2025, 38(5): 30-41.
[3] 夏正晶. “人货场”视角下元宇宙推动电商变革的路径分析[J]. 商业经济研究, 2022(20): 99-102.
[4] 战令琦. 媒介技术驱动下广告传播的新形态与新理念[J]. 青年记者, 2021(8): 11-14
[5] 毕文轩. 生成式人工智能的风险规制困境及其化解: 以ChatGPT的规制为视角[J]. 比较法研究, 2023(3): 155-172.
[6] 袁海霞, 张怡琳, 白琳, 等. 虚拟数字人对直播参与行为影响的机制及边界条件研究[J]. 管理学报, 2024, 21(12): 1849-1859.
[7] 陈力铭. 数字经济时代农产品直播电商的“信息茧房”困境及其伦理反思[J]. 西南民族大学学报(人文社会科学版), 2024, 45(7): 141-148.
[8] 郭小东. 生成式人工智能的风险及其包容性法律治理[J]. 北京理工大学学报(社会科学版), 2023, 25(6): 93-105, 117.
[9] Grootendorst, M. (2022) BERTopic: Neural Topic Modeling with a Class-Based TF-IDF Procedure. arXiv: 2203.05794.
[10] 高北晨, 张帅, 朱梅基. 全球数字共通: TikTok难民的平台迁徙与国际传播新理路——基于小红书笔记的BERTopic模型分析[J/OL]. 未来传播: 1-12. 2025-12-22.[CrossRef
[11] 徐晓瑜, 孙换, 贾庆丹. TCCM框架下直播电商研究的知识体系与趋势展望——基于混合方法的国内外文献评析[J]. 南开管理评论, 2025, 28(4): 184-195.
[12] 侯文军, 卜瑶华, 刘聪林. 虚拟数字人: 元宇宙人际交互的技术性介质[J]. 传媒, 2023(4): 25-27, 29.
[13] Giddens, A. (1984) The Constitution of Society: Outline of the Theory of Structuration. Polity Press.