Generative AI-Powered Fraud Detection in Workers’ Compensation: A DevOps-Based Multi-Cloud Architecture Leveraging, Deep Learning, and Explainable AI

Abstract

Workers’ compensation fraud is an increasingly widespread issue that creates substantial financial and social costs, significantly impacting vulnerable claimants who have legitimate claims and needs. The ongoing and persistent battle against these escalating costs necessitates the development and application of innovative new tools that can effectively address the complexities of fraud in a challenging environment characterized by diverse new and emerging delivery models. The multifaceted process of fraud detection has made notable advancements due to the strategic implementation of generative AI in various areas, thus greatly enhancing our ability to effectively combat these pressing issues. In this evolving context, fraud can be modeled with high efficacy as a generative problem, leveraging advanced deep learning techniques and explainable AI to create robust detection models that are not only efficient but also transpar- ent. To facilitate this important integration, an outline for a comprehensive DevOps-based multi-cloud workflow and architec- ture is provided, specifically aimed at incorporating Generative AI-Powered Fraud Detection strategies that are thoughtfully designed for Workers’ Compensation systems. This forward- thinking approach not only aims to substantially increase the overall efficiency of fraud detection processes but also seeks to ensure that legitimate claims are processed fairly, equitably, and without unnecessary or unjustifiable delays.