AI vs. Humans: Who do we trust?: Examining trust in AI among Swedish adults

Abstract: This study investigates trust in Artificial Intelligence (AI) among Swedish adults, focusing on how different factors influence this trust and the nuances of AI-human decision-making. A quantitative survey, distributed mainly on Swedish university campuses, centered around four scenarios: Dating, Math, Doctor and Career, aimed at gauging emotional and cognitive trust in AI and the preference for AI or human recommendations across various contexts. Analysis of responses from 195 participants brought to light several key insights. Firstly, cognitive trust in AI tended to surpass emotional trust and this varied depending on the subjectivity of the scenarios presented. Secondly, while participants often showed greater trust in human recommendations, this tendency fluctuated based on how objective they perceived the task to be. Furthermore, there was a positive correlation between experience with AI and trust, though this was not consistent across different scenarios. Lastly, a notable negative correlation was observed between technological anxiety and trust in AI. The study concludes that task nature significantly affects trust in AI, particularly in tasks seen as objective, offering important insights for AI's role in marketing and policy-making. The research recognizes demographic limitations and suggests future exploration into AI trust across diverse populations and evolving AI applications.