Why do we Trust Chatbots? From Normative Principles to Behavioral Drivers

Researchers specializing in human-computer interaction published a new study on arXiv on February 13, 2025, to investigate the psychological and behavioral drivers behind why users trust AI chatbots despite a potential lack of demonstrated reliability. The paper, titled "Why do we Trust Chatbots? From Normative Principles to Behavioral Drivers," highlights a significant disconnect between how global regulators define trust and how everyday users actually experience it during digital interactions. By analyzing the increasingly blurred boundaries between automated responses and human-like conversation, the authors seek to explain why humans are prone to over-investing confidence in these systems. The study argues that while policy frameworks and legal regulations typically approach trust from a normative perspective—focusing on transparency, accountability, and technical accuracy—actual user behavior is dictated by much more subtle cues. Instead of basing trust on a chatbot's proven ability to deliver factual information, the research suggests that users are often influenced by specific interactional design choices. These design elements, which can include friendly personas, conversational fluency, and empathetic tone, can create a false sense of security that is not necessarily earned through performance. Ultimately, the research warns that the psychological mechanisms of trust in AI are highly susceptible to manipulation through interface design. As chatbots become more sophisticated, the gap between perceived trustworthiness and actual safety risks widening. This findings suggest that future regulatory efforts must move beyond high-level ethical principles and begin addressing the behavioral triggers that lead users to trust automated systems more than they perhaps should. The study serves as a critical call for developers to prioritize genuine system reliability over superficial design features that mimic human rapport.