Advancing artificial intelligence ethics in health and genomics: lessons from a public survey in South Korea

Abstract

Advances in healthcare and genetics are becoming increasingly integrated with artificial intelligence (AI), offering transformative potential alongside complex ethical challenges. This study aimed to assess public awareness and perceptions of AI ethics in healthcare (AI-H) in South Korea, with the ultimate goal of informing the development of research ethics guidelines. A nationwide online survey was conducted from January 10 to 20, 2023, targeting the general public, and 1,002 respondents were recruited through stratified random sampling. The questionnaire explored expectations of AI-H, perceived risks, willingness to share different types of personal data, and the perceived importance of various ethical principles and education targets. A large majority of respondents (84.5%) expressed optimism about the positive impacts of AI-H over the next five years, while only 3.1% anticipated negative consequences. Key concerns included the disclosure of personal information (54.0%), potential AI errors causing harm (52.0%), and ambiguous legal responsibilities (42.2%). Willingness to share data was highest for electronic medical records (72.8%), lifestyle data (72.3%), and biometric data (71.3%), while genetic data was least preferred (64.1%). Ethical principles considered most important were privacy protection (83.9%), safety and security (83.7%), legal duties (83.4%), and responsiveness (83.3%). Developers (70.7%), medical institution managers (68.2%), and researchers (65.6%) were identified as top priorities for ethics education, whereas the general public (31.0%) and students (18.7%) ranked lower. This study represents the first nationwide assessment of public ethical awareness of AI-H in South Korea. While there is strong support for AI-H, significant concerns remain, particularly regarding data privacy and legal accountability. The findings highlight the need for expanded ethics education, especially among younger populations, and for balanced attention to ethical principles beyond privacy, such as inclusiveness and accessibility. These insights provide valuable guidance for developing socially responsible AI policies and practices in healthcare.