Singapore reinsurers warned of losses from shared AI systems
Common technology providers could expose multiple insurers to simultaneous failures.
The Singapore Reinsurers’ Association (SRA) has issued a handbook advising reinsurers to assess how artificial intelligence (AI) used by insurers could affect underwriting, pricing, claims and the risks they take on.
The October 2026 handbook focuses on AI systems used by insurers that transfer part of their risks to reinsurers, including systems operated by external technology providers.
It identifies five main risk areas: data and model reliability, third-party AI dependencies, underwriting, claims handling and concentration risk.
The association warned that reinsurers could face losses when multiple insurers rely on the same AI models, cloud infrastructure or technology providers.
A failure affecting a shared system could disrupt underwriting or claims processing across several insurers, potentially increasing losses across a reinsurer's portfolio.
The handbook also highlights risks from AI-generated pricing decisions, including inaccurate data, biased models and changes in performance over time.
For claims, it advises reinsurers to examine whether AI systems could produce incorrect settlements or affect claims costs, particularly for unusual or complex cases.
Reinsurers are encouraged to assess these risks through five stages: identifying AI exposure, assessing available evidence, deciding on appropriate measures, monitoring changes and responding to incidents.
Depending on the findings, reinsurers could adjust pricing, coverage limits, contract terms or reporting requirements.
The handbook also recommends checking whether existing insurance contracts unintentionally cover AI-related losses that were not considered when policies were priced.
The guidance complements the Monetary Authority of Singapore's existing AI governance frameworks and the MindForge AI Risk Management Toolkit.
SRA said the handbook is non-binding and does not require reinsurers to independently validate insurers' AI governance systems. It will be reviewed following the release of the regulator's final AI risk management guidelines.