Trust in AI versus Human Experts : A Developmental Perspective

Sammanfattning

As artificial intelligence (AI) systems become increasingly integrated into everyday contexts, understanding how individuals develop trust in these systems has become a central concern, particularly in developmental settings. This thesis investigated how trust in AI versus human experts is shaped by motivational context and developmental differences, while synthesizing existing research on trust in AI. A systematic review examined prior literature on child-AI interaction, focusing on developmental differences, selective trust, and contextual influences. Building on these insights, the experimental study aimed to examine whether motivational context and developmental differences influence trust in AI versus human experts. To address this, a 2 × 2 between-subjects design manipulated motivation (intrinsic vs. extrinsic) and compared children and adults (N = 180). Participants completed a gamified knowledge-based task in which they could consult either an AI system or a human expert, operationalizing trust as the proportion of AI selections. No significant effects of motivation or age groups were found. Overall, the findings suggest that trust in AI may be less influenced by short-term contextual manipulations and developmental differences, and instead more strongly grounded in stable individual attitudes.