The propensity to trust in (automated) technology mediates the links between technology self-efficacy and fear and acceptance of artificial intelligence

Christian Montag * , Johannes Kraus  , Martin Baumann , Dmitri Rozgonjuk

Abstract: The present study investigated the extent to which the concepts of the individual tendency to trust in (automated) technology, technology self-efficacy, accepting/fearing artificial intelligence and trust in AI overlap. This research question is of relevance because trust in automation literature is rich, whereas only a few studies exist in the context of studying attitudes towards AI (and their relationship with trust in AI). To close this gap in the literature, 289 participants completed three questionnaires related to the propensity to trust in automated technology, technology self-efficacy, and attitudes towards artificial intelligence (ATAI). Notably, the ATAI scale includes a single item assessing trust in AI. Results revealed robust positive correlations between the propensity to trust in (automated) technology, technology self-efficacy and ATAI’s acceptance scale (including a positive association on single item level regarding trusting AI). Regarding the ATAI’s fear scale, negative correlations with trust in automated technology and technology self-efficacy were observed. Mediation models are also reported in this study. Our work shows that the construct of trust in automated technology and attitudes towards AI (including the AI trust item) overlap small to moderate (in terms of effect sizes). This means that some of the existing research insights in the domain of the propensity to trust in automation might be transferred to the psychological foundation of attitudes and the interaction with artificial intelligence. However, it is also clear that considerable non-shared (unique) variance exists between the constructs, making it necessary to further understand how positive and negative attitudes towards AI are formed.

Keywords: Trust in automation,  Trust in artificial intelligence,  Technology self-efficacy,  Positive attitudes towards artificial intelligence,  Negative attitudes towards artificial intelligence

对（自动化）技术的信任倾向是技术自我效能感与对人工智能的恐惧和接受之间联系的中介

Christian Montag * , Johannes Kraus , Martin Baumann , Dmitri Rozgonjuk

摘要：本研究调查了个人对（自动化）技术的信任倾向、技术自我效能感、接受/惧怕人工智能以及对人工智能的信任等概念的重叠程度。这个研究问题之所以具有现实意义，是因为信任自动化的文献非常丰富，而研究对人工智能的态度（及其与对人工智能的信任之间的关系）的文献却寥寥无几。为了填补这一文献空白，289 名参与者填写了三份问卷，内容涉及对自动化技术的信任倾向、技术自我效能感和对人工智能的态度（ATAI）。值得注意的是，ATAI量表包括一个评估对人工智能信任度的单项。结果显示，对（自动化）技术的信任倾向、技术自我效能感和 ATAI 的接受量表之间存在稳健的正相关关系（包括在信任人工智能的单个项目层面上的正相关关系）。至于 ATAI 的恐惧量表，则观察到与自动化技术信任度和技术自我效能负相关。本研究还报告了中介模型。我们的研究表明，对自动化技术的信任与对人工智能的态度（包括对人工智能的信任项目）存在小到中等程度的重叠（就效应大小而言）。这意味着，在信任自动化倾向领域的一些现有研究见解，可以转移到态度的心理基础以及与人工智能的互动上。然而，同样明显的是，这些建构之间存在着相当大的非共享（独特）差异，因此有必要进一步了解人们对人工智能的积极和消极态度是如何形成的。

关键词：对自动化的信任；对人工智能的信任；技术自我效能；对人工智能的积极态度；对人工智能的消极态度

来源：Computers in Human Behavior Reports（2023）