Our own briefs on AI and digital insurance: researched from the source, checked and decided by Stet, our AI editor. Newest editions first; within an edition, in our editor’s order.
Anthropic's IPO registration statement, filed with the US SEC on 26 October, warns that a catastrophic AI incident could be larger than any single insurer, and possibly the whole industry, can underwrite. The AI developer also says the market for AI-specific cover is still young and that it may be unable to buy some AI liability cover on reasonable terms, or at all. Morningstar DBRS said the filing highlights an AI insurance protection gap.
South African digital insurance broker Naked has launched cover for cats and dogs, underwritten by Hollard, and says its AI approves eligible claims automatically while staff handle the rest. Naked's early data shows a median of 1.3 days from claim to payment, with about 60% paid within 48 hours, but it has not said how many claims this covers. Accident-only cover starts at R40 a month.
Guidewire has agreed to buy London-based hyperexponential, which sells AI underwriting and pricing software to commercial insurers, for $1.7bn in cash plus up to $400m tied to revenue targets. The core-systems vendor will fund the deal partly with $1bn of new term loans and expects to complete it in the second quarter of its 2027 fiscal year.
SpaceX has made a binding commitment to a master finance lease of about $40bn with an Apollo-managed entity, which will buy Nvidia Vera Rubin chips and related equipment and lease them to SpaceX. Apollo calls it the largest committed GPU financing. The deal is subject to closing conditions and is expected to close before the end of 2026.
Pennsylvania Governor Josh Shapiro approved Senate Bill 806, the Artificial Intelligence Deceptive Advertising Act, on 22 October as Act 99 of 2026. The act bars advertisers from using AI-generated or significantly AI-modified content in a way that gives consumers a false perception of the nature, origin or authenticity of what is advertised, unless the ad carries a clear AI disclosure. It takes effect after 60 days, on 21 December, with civil fines of up to $3,000.
Texas Attorney General Ken Paxton sued OpenAI and its chief executive Sam Altman on 23 October, alleging under the state's consumer protection law that the company misled users about ChatGPT's safety. According to Courthouse News, the 67-page suit asks the court to stop OpenAI releasing new ChatGPT models in Texas unless a court-appointed independent evaluator approves them.
A US federal appeals court on 22 October vacated a preliminary injunction against Mississippi's 2025 law on diversity programmes in public education and ordered the case reassigned to a different district judge. The Fifth Circuit panel said Judge Henry Wingate's orders contained invented parties, false quotations of the law and a case that does not exist; he had said a law clerk used AI in drafting.
Donegal Insurance Group is rolling out ZestyAI's Z-PROPERTY analytics to its underwriters and portfolio managers under a wider agreement with the vendor, FinTech Global reported on 21 October. The tool adds roof condition, roof complexity and debris indicators to roof age. The report gives no states, timetable or measured results; Donegal has used ZestyAI's Roof Age model since a partnership announced in October 2024.
US small commercial carrier MGT says independent agents can now get real quotes from it inside ChatGPT and Claude, with its AI agent Aimee checking appetite and pricing the risk. MGT calls itself the first carrier to offer chat-based quoting to agents. Every quote still opens in MGT's portal for the agent to review before binding.
Marsh chief executive John Doyle warned that broad exclusions of artificial intelligence risk in insurance policies could cause “real confusion and uncertainty”, Reuters reported. Speaking after the broker's third-quarter results, he asked for a more careful assessment of AI exposures and said that there have been very few AI claims so far.
Manulife has launched an AI assistant that answers Canadian advisors' underwriting questions before they submit an application, while decisions stay with its underwriters and its MAUDE engine. The tool draws on Manulife's underwriting handbooks and first covers life, critical illness and Synergy products. Manulife says the answers are not binding and may contain errors.
ERGO NEXT Insurance, the US small-business insurer owned by Munich Re's ERGO Group, said on 14 October that business owners can now complete commercial property inspections themselves in about 15 minutes with a vision-AI tool. The company compares this with a usual wait of five to 10 days for a field inspector and says the tool is available now for direct customers.
DXC Technology announced insurance software changes at its London executive forum on 7 October, including a new Assure Configuration Manager and wider use of its AI-enabled Assure Smart Apps. It has set out its intent to combine Assure Platform capabilities with its London Market Core Platform. DXC says Smart App deployments support more than 450 insurers.
US online insurance marketplace Insurify has joined Sierra and other technology firms as an early design partner for Poppy, a proposed AI Personal Agent Protocol. Insurify said it will push for insurance-specific rules on shopper consent, coverage details and carrier protections when AI agents request quotes for consumers.
Boston-based specialty insurer Ledgebrook raised USD 200 million in primary equity in a round co-led by Allianz X and Rockefeller Capital Management. Allianz Re also agreed to enter a multi-year reinsurance agreement with the company, which writes US excess and surplus lines cover through wholesale brokers and prices risks with its own AI platform.
The Monetary Authority of Singapore (MAS) issued Guidelines on AI Risk Management that apply to all financial institutions and all forms of AI, and says firms remain accountable for AI that third parties develop, operate or provide. The guidelines take effect on 7 October 2027, with Sections 5 and 6 due by 7 October 2028.
In a Eurofi magazine article published on 16 September, EIOPA argues that wider use of AI in finance increases dependence on a few outside providers, market concentration and cyber risk. It cites a 2025 survey in which 65% of 347 insurers in 25 EEA countries already used generative AI, and says rigorous implementation of DORA and the AI Act matters more than new AI rules.
Sapiens has launched SapiensAIP, a software-as-a-service platform that it says builds AI agents into core insurance systems for underwriting, policy, billing, claims and customer engagement. Its Migration Hub and Configuration Hub propose data mappings and system settings with confidence scores, which the insurer's own specialists review and approve. Texas-based administrator Continental General is evaluating both tools.
