AI and Risk Pooling: Evolution, Not Revolution
Recent debates, including commentary around the Mills Review, have asked whether artificial intelligence will break traditional insurance risk pools. Industry voices such as Adam Pichon of LexisNexis Risk Solutions frame the issue differently. Instead of overthrowing the foundation of pooling, AI is sharpening how insurers group risks and set prices.
Refining Underwriting Accuracy
AI improves underwriting in two complementary ways. For lines that still rely on manual assessment, machine learning speeds up document review, flags anomalies and helps underwriters prioritize cases. For highly automated product lines, models incorporate new signals from telematics, claims histories and alternative data to boost predictive power. That leads to more precise segmentation and fewer mispriced policies, which supports better capital allocation across portfolios.
The Practical Limits of Personalization
Full, individual-level pricing faces two hard constraints. First, data scarcity and privacy rules limit access to the breadth and depth of information needed for perfectly idiosyncratic pricing. Second, insurance claims contain real randomness. Even the best model cannot predict every rare event or correlated catastrophe. Overfitting to historical signals can also create fragile pricing that fails under stress. Those limits mean some degree of aggregation will remain necessary.
Towards a More Resilient Market
AI’s practical contribution is a more efficient, fairer market rather than the elimination of pooling. Better segmentation reduces cross-subsidies where appropriate, improves fraud detection and supports targeted risk mitigation programs. Regulators and market participants will still use pooling to share tail risk, maintain affordability and promote social objectives. In short, AI is a tool for refinement that bolsters market stability within long-standing insurance principles.
For investors and executives, the takeaway is strategic. Deploy AI to tighten underwriting, improve pricing accuracy and support resilience, while recognizing that fundamental risk aggregation remains a core function of insurance.
