AI-powered decision-making in facilitating insurance claim dispute resolution

Zhang, Wen, Shi, Jingwen, Wang, Xiaojun and Wynn, Henry ORCID: 0000-0002-6448-1080 (2023) AI-powered decision-making in facilitating insurance claim dispute resolution. Annals of Operations Research. ISSN 1572-9338

Text (Wynn_ai-powered-decision-making--published) - Published Version Available under License Creative Commons Attribution. Download (2MB)

• Author
• Scopus publication

Abstract

Leveraging Artificial Intelligence (AI) techniques to empower decision-making can promote social welfare by generating significant cost savings and promoting efficient utilization of public resources, besides revolutionizing commercial operations. This study investigates how AI can expedite dispute resolution in road traffic accident (RTA) insurance claims, benefiting all parties involved. Specifically, we devise and implement a disciplined AI-driven approach to derive the cost estimates and inform negotiation decision-making, compared to conventional practices that draw upon official guidance and lawyer experience. We build the investigation on 88 real-life RTA cases and detect an asymptotic relationship between the final judicial cost and the duration of the most severe injury, marked by a notable predicted R2 value of 0.527. Further, we illustrate how various AI-powered toolkits can facilitate information processing and outcome prediction: (1) how regular expression (RegEx) collates precise injury information for subsequent predictive analysis; (2) how alternative natural language processing (NLP) techniques construct predictions directly from narratives. Our proposed RegEx framework enables automated information extraction that accommodates diverse report formats; different NLP methods deliver comparable plausible performance. This research unleashes AI’s untapped potential for social good to reinvent legal-related decision-making processes, support litigation efforts, and aid in the optimization of legal resource consumption.

Item Type:	Article
Official URL:	https://www.springer.com/journal/10479
Additional Information:	© 2023 The Author(s)
Divisions:	Statistics
Subjects:	H Social Sciences > HA Statistics K Law > K Law (General) T Technology > T Technology (General)
Date Deposited:	06 Nov 2023 12:54
Last Modified:	12 Dec 2024 03:56
URI:	http://eprints.lse.ac.uk/id/eprint/120649

Actions (login required)

View Item