In light of the COVID-19 pandemic, use of virtual care technologies such as telehealth have dramatically increased. The next phase, already under development, is completely virtual AI-powered healthcare (termed ‘virtual physicians’). Therefore, we need to understand how it will be perceived by its potential users. This way, we can be prepared to address the needs and concerns of patients and inform the implementation of these technologies.
For this study, we held focus groups of students in medical and data science graduate studies to learn more about their perceptions as potential patients of virtual physicians. Many patients seek convenience and cost-effectiveness in their medical care, and our focus group participants thought that a virtual physician could offer both of these.
However, our participants had serious ethical concerns about AI-powered virtual physicians, such as the potential for discrimination (by race, ethnicity, and socioeconomic status) and accountability questions (Who is responsible for medical errors made by a virtual physician?). These should be given substantial consideration throughout the development and implementation of virtual physicians. Participants were also concerned about data privacy and the potential for misdiagnosis by virtual physicians due to incorrect algorithms or that algorithms are not updated as new knowledge comes to bear. Transparency and accountability are key during the programming and implementation of virtual physicians, and accuracy should be demonstrated using trustworthy data. Ultimately, having virtual physicians work in tandem with human physicians is likely the most optimal model.