Guest Editorial Explainable AI: Towards Fairness, Accountability, Transparency and Trust in Healthcare

Abstract

The papers in this special section focus on explainable artificial intelligence (AI) in healthcare services. Recent advances in AI, precision health, and medicine have paved the way for the accelerated adaptation and use of intelligent tools and systems in decision-making processes across the healthcare spectrum. Insights and knowledge derived from complex analytics are used to implement diagnostic and therapeutic solutions and targeted interventions in individuals and communities across the globe. Given the complexity of the current multi-dimensional clinical and public health data landscape, providing explainability in the context of socio-environmental and technical systems is a key to revealing pathways from socio-economic disadvantages to health disparities and implementing equitable interventions. As the complexity of the underlying data sets and AI-based algorithms increases, the explainability and justifiability of the insights generated decrease. Humans need to understand the underlying mechanism behind these insights to know whether they are sound, correct, trustable, and justifiable to make informed decisions. Lack of understandability and explainability in the biomedical domain often leads to poor transparency and accountability and ultimately lower quality of care and suboptimal and unfair health policies. Explainability is considered one of the prerequisites for deep medicine, where AI is meant to provide composite, panoramic views of individuals’ medical data.