Explainable AI for High-Stakes Decision Making in Healthcare

EasyChair Preprint no. 13895

Abstract

This research explores the development and implementation of explainable artificial intelligence (AI) models in healthcare, focusing on delivering accurate diagnoses and treatment recommendations with transparent and understandable reasoning for medical professionals. Explainable AI aims to bridge the gap between advanced computational models and the practical needs of healthcare providers by making AI-driven decisions interpretable and trustworthy. By providing clear explanations of AI reasoning, these models can enhance clinical decision-making, increase trust in AI systems, and improve patient outcomes. This study highlights the critical need for explainability in high-stakes healthcare settings, where understanding the rationale behind AI decisions is essential for gaining acceptance among medical professionals and ensuring patient safety. Furthermore, the research examines various techniques for achieving explainability, such as visualizations, natural language explanations, and rule-based systems, and evaluates their effectiveness in clinical applications. The goal is to promote the integration of explainable AI in healthcare, thereby fostering transparency, accountability, and ultimately, better healthcare delivery.

Keyphrases: AI systems, AI-driven diagnoses, clinical decision making, Explainable AI, Healthcare, interpretability, Medical Professionals, Natural Language Explanations, Patient Outcomes, patient safety, rule-based systems, transparency, treatment recommendations, Trust, visualization

@Booklet{EasyChair:13895,
  author = {Dylan Stilinki and Joseph Oluwaseyi},
  title = {Explainable AI for High-Stakes Decision Making in Healthcare},
  howpublished = {EasyChair Preprint no. 13895},

  year = {EasyChair, 2024}}