Bhattacharya Sukriti, Castignani German, Masello Leandro, Sheehan Barry
Human-Centered AI, Data and Software (HANDS), Luxembough Institute of Science and Technology, Maison de l'innovation, Esch-sur-Alzette, Luxembourg.
Kemmy Business School, University of Limerick, Limerick, Ireland.
Front Artif Intell. 2025 Apr 9;8:1568266. doi: 10.3389/frai.2025.1568266. eCollection 2025.
This paper comprehensively reviews artificial intelligence (AI) applications in the insurance industry. We focus on the automotive, health, and property insurance domains. To conduct this study, we followed the PRISMA guidelines for systematic reviews. This rigorous methodology allowed us to examine recent academic research and industry practices thoroughly. This study also identifies several key challenges that must be addressed to mitigate operational and underwriting risks, including data quality issues that could lead to biased risk assessments, regulatory compliance requirements for risk governance, ethical considerations in automated decision-making, and the need for explainable AI systems to ensure transparent risk evaluation and pricing models. This review highlights important research gaps by comparing academic studies with real-world industry implementations. It also explores emerging areas where AI can improve efficiency and drive innovation in the insurance sector. The insights gained from this work provide valuable guidance for researchers, policymakers, and insurance industry practitioners.
本文全面回顾了人工智能（AI）在保险业中的应用。我们重点关注汽车保险、健康保险和财产保险领域。为开展本研究，我们遵循了系统评价的PRISMA指南。这种严谨的方法使我们能够全面审视近期的学术研究和行业实践。本研究还确定了为减轻运营和承保风险必须解决的几个关键挑战，包括可能导致有偏差风险评估的数据质量问题、风险治理的监管合规要求、自动化决策中的伦理考量，以及对可解释人工智能系统的需求，以确保透明的风险评估和定价模型。通过将学术研究与实际行业实施进行比较，本综述突出了重要的研究差距。它还探索了人工智能可以提高效率并推动保险行业创新的新兴领域。从这项工作中获得的见解为研究人员、政策制定者和保险行业从业者提供了宝贵的指导。
