Trust in AI Revisited – The Enduring Role of Propensity to Trust

Abstract

The importance of trust in artificial intelligence (AI) continues to grow, as trust is widely regarded as a critical prerequisite for organizational AI adoption. In this context, intention to use AI can be understood as a consequence of the decision to trust AI and is therefore strongly influenced by trust. Moreover, trust is regarded as essential for understanding the impact of increasing interaction with AI systems on both individuals and society. Much of the discussion on trust in AI relies on frameworks derived from trust in automation, but these approaches remain largely theoretical and insufficiently validated. One important empirical contribution addressing this gap is the path model developed by Karg, Ritz and Asprion (2025), which examined trust in ChatGPT using a student sample. This model conceptualizes perceived trustworthiness through performance, process, and purpose. Together with a user’s propensity to trust, these factors are assumed to determine trust in AI. Karg, Ritz and Asprion (2025) demonstrated that perceived trustworthiness is significantly shaped by users’ inherent propensity to trust, in turn, influences the intention to use AI. The present study replicates this path model using a business sample to assess the robustness of the original findings and to advance theory building. An online survey was conducted among 97 employees of a major Swiss bank, employing identical items and methodologies as in the original study. The replication largely supports the original findings. However, in contrast to the original study, performance did not significantly predict trust in AI in the business sample. The findings further reinforce the argument that users’ dispositional characteristics may play a more decisive role in shaping perceived trustworthiness of and trust in AI systems.