Trust and Technology Acceptance: Comparing Traditional Search Engines and Artificial Intelligence for Colorectal Cancer Information Seeking

Abstract

Introduction: The integration of artificial intelligence (AI) into health information seeking is transforming health promotion. Understanding how users accept and trust these communication technologies is critical for health communication and cancer control. This study examined how the Technology Acceptance Model II (TAM II) applies to colorectal cancer information seeking, comparing link-based search (e.g., Google search) versus generative response paradigms (e.g., ChatGPT/AI) while examining trust, perceived threat, and contextual factors in technology use decisions. Methods: A prospective, randomized 2×2 factorial experiment was conducted with 764 Texas adults randomly assigned to conditions to view either Google search results or ChatGPT responses for colorectal cancer symptoms, presented in either high-concern or low-concern scenarios. Participants completed validated measures including TAM II constructs adapted from Davis (1989) and Kamal et al (2020), multidimensional trust scales, Extended Parallel Process Model threat measures (Witte, 1992), and technology-related stress items, all demonstrating acceptable reliability (α > .77). Data analysis included two-way ANOVAs, correlation analysis, and stepwise regression modeling. Results: Google search received significantly higher ratings than AI across all Technology Acceptance Model II constructs. Technology preferences appeared to reflect multiple factors including interface familiarity, trust in information sources, and usability expectations, with traditional search benefiting from established user mental models and transparent source attribution. Trust emerged as the strongest predictor of behavioral intention. No significant main effects were found for concern level, and no interaction effects emerged between technology type and concern level, indicating that technology preferences remained consistent regardless of symptom severity. Conclusions: For cancer control and prevention, these findings suggest that patients seeking colorectal cancer symptom information may be more likely to trust and act upon traditional search results than AI-generated responses, focusing on technology use intentions for health information seeking that directly inform cancer screening and care-seeking behaviors, potentially affecting screening behaviors and care-seeking timing. Current AI implementations may not optimally serve health information needs with lower acceptance potentially related to limited source transparency and increased cognitive demands compared to familiar search interfaces, as suggested by preference patterns. Cancer control professionals should anticipate that the growing integration of AI into health information seeking may influence the public’s cancer symptom evaluation and screening behaviors.