AS autonomous “agentic AI” systems increasingly manage business decisions, insurers are grappling with how to cover the risks these technologies create. Advanced AI is designed to replace much of human supervision, yet its errors — such as “hallucinations” where fabricated data is presented as fact — expose companies to new liabilities.
Traditionally, insurers addressed such risks through “silent coverage,” meaning AI-related losses were implicitly included under general policies. But experts like Sonal Madhok and Anat Lior compare today’s situation to the early cyber-insurance era, predicting that policies will soon explicitly define AI coverage.
Insurers are responding in varied ways. Some, like Armil la, test AI models for vulnerabilities and offer specialized plans, though they exclude high-risk areas such as medical diagnostics. Others, such as Founder Shield, now include “AI malfunction and hallucination” scenarios in professional service insurance, even covering real-world mistakes like erroneous inventory orders. Conversely, some providers insert “absolute AI exclusion” clauses to deny coverage altogether.
Global insurer Munich Re also underwrites AI developers and users, acknowledging that model errors cannot be fully eradicated. Analysts foresee huge potential in the emerging market — Deloitte projects AI insurance premiums could reach $4.8 billion globally by 2032.
AFP
#TheGlobalNewLightOfMyanmar