Children and adolescents' views on artificial intelligence in pediatric healthcare: a qualitative focus group study

The increased use of artificial intelligence (AI) in healthcare, including pediatrics, offers significant opportunities for improving quality, personalization, and effectiveness of care. Nevertheless, successful implementation of this approach in pediatric patients depends on their acceptance and the integration of their preferences regarding AI use within participatory decision-making. Since previous research insufficiently addresses the perspectives of children and adolescents on AI use in pediatric healthcare, the objective of this study was to address this research gap.

This qualitative study was conducted at the University Children's Hospital Duesseldorf, Germany, and in surrounding schools. Recruitment took place from March to September 2024. Semi-structured interviews including two scenarios were conducted: an AI-based application (1) for assessing the development of children and adolescents, (2) for supporting and monitoring children and adolescents with chronic conditions. Content analysis was used to evaluate information.

Eighteen focus group interviews with 58 participants (8-11 years = 10; 12-14 years = 23; 15-17 years = 25) were conducted. Nine key topics emerged: benefits, fears, dealing with misdiagnoses of AI, responsibility for data input and access, impact on the physician-patient relationship, the right not to know, transparency, data sources, and desired features. Participants repeatedly emphasized that AI should serve only as a supportive tool, with healthcare professionals retaining final authority and critically evaluating its findings. Key benefits included early detection of health anomalies and reduced diagnostic or human errors. Barriers to acceptance encompassed data privacy concerns, suspected lack of empathy, absence of physical examination, and risk of false diagnoses.

This qualitative study underlines, that the requirements, perceived benefits, and concerns of AI-based applications in pediatric healthcare are multifaceted and necessitate including minors in their development and implementation. Age-appropriate education on AI use, risks, limitations, opportunities and strengths is essential for acceptance. If possible, minors should be openly informed about the limitations of AI decision transparency, with underlying mechanisms explained where appropriate. At best, children and adolescents should already be involved in development of AI applications in terms of information and consent.

This study was retrospectively registered in the German Clinical Trials Registry (DRKS00034474; 17th June 2024).