Delving into AI disclosure architecture: Building trust in Indonesian telemedicine advertising

Type of the article: Research ArticleAbstractAI disclosure has become increasingly important in digital advertising, particularly in high-risk services such as telemedicine, where consumers evaluate message quality alongside provider accountability. This study aimed to examine how AI disclosure architecture, reflected in disclosure timing, disclosure specificity, and human control, influences perceived authenticity, message credibility, brand trust, and usage intention in Indonesian telemedicine advertising. A quantitative scenario-based online experiment used a 2 × 2 × 2 between-subjects design with 527 valid respondents. Data were analyzed using three-way factorial ANOVA, confirmatory factor analysis, structural equation modelling, and bootstrap analysis with 5,000 resamples. Early disclosure, high specificity, and stated professional review or oversight produced more favourable evaluations across the disclosure conditions examined. Human control generated the largest main effects on perceived authenticity (partial η² = 0.346), message credibility (partial η² = 0.316), brand trust (partial η² = 0.318), and usage intention (partial η² = 0.166). Perceived authenticity increased brand trust (β = 0.304, p < 0.001), message credibility had a stronger effect on brand trust (β = 0.598, p < 0.001), and brand trust increased usage intention (β = 0.736, p < 0.001). Timing and specificity showed significant indirect effects on brand trust, while human control showed both direct and indirect effects. The effects were predominantly additive, with only a small timing × human-control interaction for message credibility (partial η² = 0.009). These findings indicate that, when AI disclosure is present, timing, specificity, and stated professional oversight provide distinct trust-relevant information rather than functioning as a strongly synergistic configuration.Acknowledgment(s)The authors would like to express their sincere gratitude to all respondents who participated in this study and provided valuable responses for the experimental survey. The authors also thank the academic colleagues and research reviewers who provided constructive feedback during the development of the research instrument, experimental stimuli, and manuscript refinement. Their insights helped improve the clarity, methodological rigor, and overall quality of this study. Any remaining errors or deficiencies are solely the responsibility of the authors.