The use of artificial intelligence (AI) for self-diagnosis is widespread and continues to grow.Effective adoption of AI depends not only on its accuracy but also on patient trust.This study examines trust in AI healthcare platforms, focusing on two key antecedents: AI benevolence (empathy) and competence.We propose a model in which both factors influence trust, with the effect of competence moderated by task complexity.A quasi-experimental methodology is planned to test how users' perceptions of