Manipulating Trust of Autonomous Products With Affective Priming

Abstract

Abstract User-to-product trust has two notable aspects: (1) the user’s propensity to trust, and (2) the product’s trustworthiness as assessed by the user. Autonomous products, which perform many functions on their own with limited user input, require the user to exhibit trust at an appropriate level before use. Research in product trust thus far has focused on the product trustworthiness: manipulating the product’s design, for example, anthropomorphizing an autonomous vehicle and measuring changes in trust. This study flips the usual approach, manipulating a person’s propensity to trust and measuring response to an existing autonomous product, the Amazon Echo. We build on our past successes with priming exercises to reveal insights into the user-related factors of product trust. In this study, we used visual stimuli that evoked either positive, neutral or negative emotions as affective primes to influence users’ trust propensity before the interaction. The participants interacted with a mock-up of the Amazon Echo via ten pre-determined question-and-answer (Q&A) sets. During the interaction, the participants evaluated the Echo’s competence and if it met participants’ expectations. They also reported trust towards the Echo after the Q&A sets. Holistically, the affective primes show no significant effect on the trust propensity. For the subgroup of participants whose expectations of the product’s performance were met, both the perceived product competence and the affective primes have significant effects on trust propensity. These results demonstrate the complex nature of trust as a multidimensional construct and the critical role of product performance in trust formation. They also suggest that it will be difficult for a product to build trust with users who expect the product to perform in a different way than its intent — if one wants to design a product that builds trust, they should understand user expectations and design to meet them. This learning can facilitate the intentional design of the affective process in trust formation that helps build a healthy level of trust with autonomous products.