AI and the Transformation of General Insurance: Risk, Pricing and Governance
Abstract
Artificial intelligence (AI) is reshaping the architecture of general insurance by altering how insurers gather information, assess risk, select and price exposures, manage claims and detect fraud. Existing scholarship tends to approach these developments either through actuarial modelling, which treats AI primarily as an improved estimator, or through general accounts of responsible AI, which treat insurance as one application domain among many. This article develops an integrated technical and legal framework for analysing AI across general-insurance lines. It advances six analytical propositions (P1–P6), a dual-channel account of technical and governance effects, and a decision-role ladder distinguishing advisory, augmented and determinative uses of AI. Comparative analysis across motor, property, travel, liability, cyber, professional indemnity, marine and aviation insurance shows that AI does not have a uniform pricing or risk effect: in some lines it sharpens measurement of recognisable risks, while in others it changes causal structures, responsibility and correlation. The framework is further illustrated through publicly documented AI practices at Allianz and Munich Re, used as non-causal corporate illustrations rather than evidence of proprietary model performance. The analysis argues that AI can improve the insurer’s information position while simultaneously relocating uncertainty and creating new model, data, explainability, oversight and accountability exposures. Governance should therefore track the decision role and material consequences of AI use rather than technical sophistication alone.