Read up on September's leading AI and insurance stories driving industry momentum, regulatory progress, and business change.
Latest Articles as of September 3
News: US pushes G20 to skip new AI rules
The brief: The US pressed G20 members on Tuesday to take a hands-off approach to AI regulation at a two-day innovation ministerial in Chapel Hill, North Carolina, co-hosted by Commerce Secretary, Howard Lutnick, and tech adviser, Michael Kratsios. Kratsios urged countries to adopt "Carolina Principles," reserving new rules for novel considerations rather than treating AI as its own category, and said China had signed on. Elon Musk criticized the EU's tech regulations as inhibiting progress, while Mark Zuckerberg and Demis Hassabis were also slated to address the ministers by video. The push comes as the US races China for AI supremacy, with a UN panel recently warning that AI development is outpacing both scientific understanding and government policy.
News: Cyber premium holds steady as AI reshapes the threat picture
The brief: Global cyber insurance premium is on track to reach $16.4 billion in 2026, growing at a steady 5% annual clip since 2022, even as AI reshapes the threat landscape, according to Swiss Re's latest cyber reinsurance report. Rather than creating entirely new categories of loss, the report argues AI is amplifying cyber risks insurers already know, since AI models often fall within existing policy definitions of computer systems. Rates fell for a fourth year, though the size of the decline eased from about 13% in 2025 to 5% in 2026 globally. North America still holds two-thirds of global premium at $10.7 billion, but Swiss Re flagged a limit-adequacy problem, with large-corporate US buyers averaging $120 million in purchased limits against a claims database showing 10 losses a year exceed that benchmark.
News: AI reshapes how people shop for insurance
The brief: Nearly three in 10 auto and home insurance customers have already used AI to research coverage, shop for policies, service accounts, or better understand their policies, according to a recent JD Power study. Among those who used AI to research products and coverage, 37% changed their policy based on what they learned and 42% purchased a policy after using AI-assisted shopping tools. Customers were least comfortable using AI for claims and policy changes, where the stakes of an error are higher, and some who began with an AI tool ultimately contacted a human representative anyway. The findings point to agents shifting from information providers toward trusted advisors as AI absorbs more of the early research.
News: Fed and FSB deliver matching AI cyber warning within days
The brief: The Federal Reserve's Spring 2026 Financial Stability Report found 50% of surveyed market participants cited AI as a major risk to US financial stability, up sharply from 30% six months earlier. Days later, the Financial Stability Board Chair told G20 finance ministers that frontier AI's impact on cyber risk is the single most immediate threat to the global financial system, capable of changing the speed, scale, and economics of an attack given the sector's dependence on a small number of concentrated cloud and technology providers. Hartford's Adrien Robinson told Insurance Business that cyber rates look "a little disconnected" from the underlying risk trajectory.
News: AI chatbots skip market-share leaders when recommending insurers
The brief: A Q2 2026 Insurance AIVI Benchmark ran 300 consumer-style insurance shopping prompts across ChatGPT, Gemini, Perplexity, Copilot, and Claude, generating 1,500 responses with 4,328 organic mentions of 134 insurance brands. GEICO led in auto, State Farm in home and renters, MassMutual in life, and The Hartford in commercial. The central finding is that being mentioned and being recommended are different measures, with The Hartford appearing in the top three at an 88.1% rate but earning an active recommendation in only 27.6% of those appearances, the lowest among the top ten brands. Market share offered little predictive value for AI visibility, with Amica's share of AI-generated home-insurance visibility running 10.8 times higher than its NAIC-reported market share.
News: Clinical AI could become a malpractice defense asset
The brief: The CEO of a medical malpractice insurtech told Insurance Business that no carrier can price AI risk with precision yet, since malpractice has a long reporting tail and clinical AI has only been in widespread use about two years. The CEO pointed to Waymo's published safety data, which shows 82% fewer injury-causing crashes than the human benchmark across 220.6 million rider-only miles, as the reason the CEO expects clinical AI to reduce medical error frequency rather than increase it. The CEO's advice to risk managers is to inventory every AI tool in clinical use, document clinical reasoning alongside tool outputs, and review vendor contracts for clinical-responsibility disclaimers before a claim arises.
