算法还是专家?任务类型对人机信任和价值感知的影响

Algorithm or Human Expert?Effects of Task Type on Human-computer Trust and Perception of Value

目的探究在人机交互过程中,任务类型对个体对算法和人类专家的认知信任及价值感知的影响。方法运用单因素两水平(主观任务和客观任务)的被试间实验设计,采用行为决策研究中的滴定法范式定量评估对算法的相对价值感知,并通过问卷测量个体对算法和人类专家的认知信任。结果在主观任务中,个体对人类专家在信心、准确率和价值感知方面都显著高于算法;但在客观任务中,个体对算法和人类专家的信心及价值感知并无显著差异,仅在准确率这一维度上,算法高于人类专家;此外,个体在进行客观任务时对算法的价值感知高于在进行主观任务时的价值感知。结论任务类型对人机信任和价值感知有重要影响,个体在主观任务下表现出了“算法拒绝”。

The work aims to explore how task type affects people's cognitive trust and value perception of algorithm and human expert judgement during human-computer interaction.The experiment applied titration paradigm which was used in decision-making domain to test the perception of advisors'value in different conditions of tasks(subjective vs.objective tasks).And the cognitive trust was measured by the questionnaire method.It was found that in the subjective task,individuals'confidence,perception of accuracy and value from human experts were higher than those of algorithms.But in the objective task,there were no observable differences on individuals'confidence and perception of value between the two advisors(algorithms vs.human experts).Individuals only perceived higher accuracy from algorithms over human experts.Last,individuals'perception of value towards algorithm was higher in the objective task than in the subjective task,Thus,task type has significant effects on individuals'trust and perception of value towards algorithms and human experts.Individuals are averse to algorithms in the subjective task compared with human experts.