Resolving insurance claims with Artificial Intelligence powered decision making

Abstract

This study explores how Artificial Intelligence (AI) can improve the speed and efficiency of dispute resolution in road traffic accident (RTA) insurance claims, which can benefit the society and the economy. We propose and apply a systematic AI-based method to estimate the costs and guide the negotiation process, instead of relying on official guidelines and lawyer expertise. We use 88 real-life RTA cases and find a strong correlation between the final judicial cost and the length of the most severe injury, with a high predicted R2 value of 0.527. We also demonstrate how various AI tools can help with information extraction and outcome prediction: · How regular expression (regex) can obtain accurate injury data for further analysis; · How different natural language processing (NLP) techniques can make predictions directly from text. Our RegEx framework can automatically extract information from different report formats; different NLP methods can produce similar reasonable results. This research shows how AI can be used for social good to transform legal-related decision-making processes, support legal actions, and optimize legal resource use.