财经论丛 ›› 2026, Vol. 42 ›› Issue (7): 88-99.

• 金融与投资 • 上一篇    下一篇

生成式人工智能应用与居民数字金融行为——来自江西省数字支付普惠调查的微观证据

习明明, 毛鹏飞, 张卢千漪   

  1. 江西财经大学数字经济学院,江西 南昌 330013
  • 收稿日期:2025-10-16 出版日期:2026-07-10 发布日期:2026-07-09
  • 作者简介:习明明(1983—),男,江西新余人,江西财经大学数字经济学院教授;毛鹏飞(2002—),男,江西赣州人,江西财经大学数字经济学院硕士生;张卢千漪(2000—),女,江西吉安人,江西财经大学数字经济学院博士生。
  • 基金资助:
    国家社会科学基金重大项目(25&ZD172);国家自然科学基金地区项目(72264011);江西省社会科学基金重点项目(24ZXQH06)

Generative Artificial Intelligence Application and Residents' Digital Financial Behavior: Micro Evidence from the Survey on Inclusive Digital Payments in Jiangxi Province

XI Mingming, MAO Pengfei, ZHANG Luqianyi   

  1. School of Digital Economics, Jiangxi University of Finance and Economics, Nanchang 330013, China
  • Received:2025-10-16 Online:2026-07-10 Published:2026-07-09

摘要: 利用2025年江西省数字支付普惠调查数据,从数字平台理财产品投资的视角实证检验了生成式人工智能应用对居民数字金融行为的影响及作用机制。研究发现,生成式人工智能应用显著促进居民参与数字金融。机制分析表明,生成式人工智能应用通过提振未来预期、重构心理账户和改变风险偏好,促进居民参与数字金融。异质性分析表明,生成式人工智能应用的该促进作用在低学历、非正规部门从业以及中老年群体中更为显著,呈现出“数字鸿沟”的“弥合效应”。进一步分析发现,生成式人工智能应用在促进居民参与数字金融后,能够显著提升弱势群体和中老年群体的福利水平。研究结果为理解人工智能技术在数字金融领域的应用提供了新的经验证据,对于推动我国数字金融高质量发展具有重要的理论价值与现实启示。

关键词: 生成式人工智能, 数字金融行为, 信息过载, 认知偏差, 数字鸿沟

Abstract: Based on data from the Inclusive Digital Payment Survey (IDPS) of 2025 in Jiangxi Province, this paper empirically examines the impact and underlying mechanism of the application of Generative Artificial Intelligence (GenAI) on residents' digital financial behavior, specifically from the perspective of digital platform on wealth management product investment. The study finds that the application of GenAI significantly promotes residents' participation in digital finance. Mechanism analysis indicates that GenAI fosters resident participation by boosting future expectations, reshaping mental accounts, and altering risk preferences. Heterogeneity analysis reveals that the promoting effect is significantly stronger among individuals with lower education levels, people with informal sector employment, and older-age groups, demonstrating its bridging effect on the digital divide. Further analysis suggests that, besides promoting resident participation, the application of GenAI significantly enhances the welfare levels of vulnerable and older groups. The research findings provide novel empirical evidence for comprehending the application of artificial-intelligence technologies in digital finance, thereby endowing both theoretical significance and practical implications for advancing the high-quality development of China's digital-finance sector.

Key words: Generative Artificial Intelligence, Digital Financial Behavior, Information Overload, Cognitive Bias, Digital Divide

中图分类号: