Artificial Intelligence in Healthcare

Abstract

Abstract Artificial intelligence (AI) is poised to significantly impact healthcare systems, including clinical diagnosis, healthcare administration and delivery, and public health infrastructures. In the context of the Quintuple Aim of healthcare (patient outcomes, cost reduction, population impact, provider wellness, and equity/inclusion), this chapter discusses the current state of AI in healthcare, focusing on issues that may inform the development of adaptive, efficient, and equitable governance frameworks for AI in healthcare. The chapter introduces prominent examples of clinical AI applications in recent years, highlighting their successes and extant limitations. It emphasizes the processes of clinical AI algorithm development, implementation, and provider adoption, noting important policy considerations for active maintenance and updating of such algorithms. It also focuses on the issue of bias in AI algorithms for healthcare by (1) illustrating how unrepresentative and/or inappropriate datasets can exacerbate health disparities and inequities, and (2) emphasizing the need for diversity, transparency, and accountability in algorithm development. It provides an overview of current national and international regulatory approaches for AI-driven medical devices. It concludes with recommendations of strategic goals for developers, healthcare providers, and governmental agencies to work towards cooperatively in building a productive and equitable future for AI in healthcare.