Trust and Reliance on AI-Generated Advice: Investigating User Perceptions and Behavioral Dynamics in Conversational AI

Abstract

The rapid proliferation of conversational artificial intelligence (AI) tools, such as ChatGPT and Gemini, has transformed the way individuals seek guidance, information, and problem-solving assistance. While these tools offer unprecedented convenience and accessibility, questions remain regarding how users perceive and trust AI-generated advice, and the potential implications for reliance, critical evaluation, and decision-making. This study investigates adults’ trust in AI-generated answers, examining both cognitive and emotional determinants, usage patterns, and behavioral outcomes. A mixed-method design was employed, involving a survey of 30 adults to assess trust, perceived accuracy, reliance, and critical evaluation, alongside interviews and a focus group with 10 adults to gain qualitative insights into personal experiences with conversational AI. Quantitative findings indicate that trust is positively correlated with perceived accuracy, frequency of use, and reliance on AI for personal decisions, while critical evaluation remains relatively low. Factor analysis identified three key dimensions of trust: perceived competence, emotional assurance, and convenience. Qualitative results revealed that participants value AI’s non-judgmental tone and ease of access, yet express cautious skepticism regarding accuracy in complex or sensitive contexts. The study highlights that trust in AI is multidimensional, shaped by cognitive, emotional, and practical considerations, and functions as both an enabler and potential constraint. Users’ reliance on AI facilitates efficiency but may risk over-dependence and reduced critical thinking. Findings underscore the importance of AI literacy, system transparency, and responsible design to promote informed trust and balanced use. These insights contribute to understanding how humans interact with AI-generated advice and offer guidance for developers, educators, and policymakers seeking to foster responsible engagement with AI technologies.