Exploring the social dimensions of AI integration in healthcare: a qualitative study of stakeholder views on challenges and opportunities.

Abstract

This study aimed to investigate the opportunities and challenges associated with integrating artificial intelligence (AI) in healthcare by exploring the perspectives of various stakeholders. The objective was to provide a nuanced understanding of stakeholder views to address concerns and promote the acceptance and successful integration of AI technologies in medical practice. This exploratory qualitative study used semi-structured interviews. Data were analysed using a combination of deductive and inductive coding, followed by content analysis to identify and develop categories. This study was conducted in Tübingen, Germany, within the framework of the TüKITZMed project (Tübingen AI Training Center for Medicine), between August 2022 and March 2023. A total of 38 stakeholders participated, including 6 lecturers, 9 clinicians, 10 healthcare students, 6 AI experts and 7 institutional stakeholders. Inclusion criteria included professionals involved in or affected by AI in healthcare, while exclusion criteria comprised individuals without relevant experience. Not applicable. The main outcome was the identification of thematic categories capturing stakeholders' perceptions, expectations and concerns regarding the integration of AI in healthcare. The analysis identified two main thematic categories: two main categories encompassing a total of 14 subcategories: (1) perceived opportunities of AI in medicine, including aspects of increased efficiency, reduced workload and improved patient safety and (2) perceived challenges of AI in medicine, such as its impact on medical decision-making and concerns about dependence on technology. These themes reflect diverse perspectives and insights across stakeholder groups. Diverse stakeholder perspectives offer valuable insights into the anticipated effects of AI in healthcare. Understanding these perspectives can support decision-makers in designing context-sensitive AI strategies and identifying areas for further professional and institutional development. Future research should monitor how these attitudes evolve in response to technological progress and real-world implementation.