算法厌恶:理论框架及研究展望

Algorithm Aversion:Theoretical framework and research prospects

研究表明,尽管算法通常比人类表现更好,在算法与人类之间,人们还是选择人类提供服务,这种现象称为算法厌恶,得到了管理学、心理学等研究领域的高度关注。然而,关于算法厌恶的实证研究对于算法厌恶产生的前因和内在机制尚无定论,部分研究发现了算法欣赏的相反结果。通过对算法厌恶的实证研究进行文献回顾,本研究归纳与总结了影响算法厌恶的算法特征、个体特征和任务特征,算法厌恶产生的认知机制和情感机制,以及深层的理论解释。在此基础上,本研究从算法厌恶的边界条件扩充、内在机制挖掘、干预策略研究和引发后果探索四个方面为未来的研究提供了建议。

Research shows that algorithms usually outperform human.Yet when people are deciding whether to use a human agent or an algorithm,they often choose the human agent.This phenomenon,which we call algorithm aversion,has drawn intense attention from management,psychology and other research fields.However,empirical studies on the antecedents and mechanisms of algorithm aversion are inconclusive and papers find the opposite effect of algorithm appreciation.Through a literature review of the empirical studies,this study shows the effects of algorithm characteristics,individual characteristics and task characteristics on algorithm aversion,and the cognitive and affective mechanisms as well as the theoretical explanation underlying algorithm aversion.On this basis,we provide suggestions for future research from four aspects:the boundary conditions,the internal mechanism,the intervention strategies and the consequences of algorithm aversion.