Artificial intelligence must evolve rapidly to quantify its own risk exposures—and thereby attract reinsurance capital, even as the actuarial models behind those calculations are themselves being reshaped by AI, Kevin Kalinich, Intangible Assets Global Collaboration Leader at Aon, told a U.S. Senate Banking Subcommittee yesterday.
Industry executives and regulators had gathered to examine AI’s expanding role in insurance and capital markets. Kalinich warned that today’s actuarial tools “lack sufficient historical loss data” to price emerging AI exposures such as deepfakes under crime policies and hallucination-driven errors under E&O programs.
Because of those blind spots, carriers are already “carving out exclusions, narrowing sub-limits, or imposing extra premiums tied to governance controls,” he noted. The process echoes early cyber-insurance development: “We’re seeing the market do exactly what it did in early cyber: carve-outs first, then limited buckets of capacity, and only later broader limits once we have data.”
Lawmakers pressed witnesses on whether the reinsurance market can scale as fast as risk evolves. Kalinich again pointed to cyber’s learning curve:
