This study applies protection motivation theory (PMT) to the COVID-19 contact-tracing context by including privacy concerns, collective efficacy, and a mediator (fear of COVID-19) and tests it in the US and South Korea. The study uses a structural equation modeling (SEM) approach and a sample of 418 Americans and 444 South Koreans. According to the results, fear was positively associated with adoption intentions in the US sample but not in the Korean sample. Coping appraisals positively affected adoption intentions in both samples. However, while all types of coping appraisals were significant in the Korean sample, response efficacy was the only significant predictor among US participants. Privacy concerns were negatively associated with adoption intentions in the US sample, but not in the Korean sample. The results indicate that differences exist in the mediating role of fear connecting threat appraisals and adoption intentions in both countries. These findings hold important implications for future studies in AI-based health communication, especially in the areas of privacy management, protection motivation, and diverse cultural contexts.