AI and perception biases in investments: An experimental study

• 成果类型: Article
• 署名作者: Fedyk, Anastassia; Kakhbod, Ali; Li, Peiyao; Malmendier, Ulrike
• 署名单位: University of California System; University of California Berkeley; National Bureau of Economic Research; Center for Economic & Policy Research (CEPR)
• 刊物名称: JOURNAL OF FINANCIAL ECONOMICS
• ISSN/ISSBN: 0304-405X
• DOI: 10.1016/j.jfineco.2026.104350
• 发表日期: 2026-11
• 页码: 104350
• 关键词: large language models; Behavioral biases; Experimental economics; Investment preferences; Financial surveys; generative AI

摘要

AI promises to accelerate and broaden access to automated investment advice. But can it capture the investment preferences and rationales of historically underrepresented investors? We ask 1272 human survey respondents and 1350 AI-generated agents to rate stocks, bonds, and cash. First, default AI-generated responses overrepresent young, high-income individuals. However, algorithmic bias is reduced with demographically-seeded prompts. Second, AI-generated free-form responses correctly reflect human rationales: risk and return, financial knowledge, and past experiences. Third, AI can help identify where a lack of financial knowledge poses issues in human responses, as shown in textual analyses of transitivity violations.