AI视角下建议采纳中的算法厌恶——“认知–社会–技术”三维动态模型

Algorithm Aversion in Recommendation Adoption from the Perspective of AI—A Three-Dimensional Dynamic Model of “Cognition-Society-Technology”

摘要: 本文探讨了人工智能视角下算法厌恶的现象，构建“认知–社会–技术”三维动态模型解析其生成机制发现；在认知层面，人们常高估自身能力，当算法削弱其决策控制感时，会通过降低算法权重来维持心理平衡；在社会层面，人与机器的情感联结缺失，以及群体压力显著加剧不信任；在技术层面，算法解释不清晰或过度复杂、运行不稳定会直接降低接受度。研究提出通过认知训练、增强人机情感互动和优化技术设计来改善问题，并指出需重点关注技术适应、建立人机协作评估标准等未来方向。

Abstract: This article explores the phenomenon of algorithm aversion from the perspective of artificial intelligence, and constructs a three-dimensional dynamic model of “cognition - society - technology” to analyze its generation mechanism; At the cognitive level, people often overestimate their own abilities. When algorithms weaken their sense of decision control, they maintain psychological balance by reducing the weight of the algorithm; At the societal level, there is a lack of emotional connection between humans and machines, and group pressure significantly exacerbates distrust; At the technical level, unclear or overly complex algorithm explanations, as well as unstable operation, can directly reduce acceptance. The study proposes to improve the problem through cognitive training, enhanced human-machine emotional interaction, and optimized technology design, and points out the need to focus on future directions such as elderly technology adaptation and establishing human-machine collaboration evaluation standards.