Here's your roundup on the latest AI and insurance news moving markets, regulators, and operations this month.
News: AI transforming catastrophe risk modeling
The brief: A principal actuary at Milliman says AI advancements are rapidly reshaping catastrophe risk modeling in insurance. Firms move from traditional actuarial methods to more advanced machine learning techniques, such as gradient boosted and explainable boosting models. These newer approaches improve risk assessment accuracy while maintaining the transparency required for regulatory approval. Adoption varies across insurers due to infrastructure limitations and regulatory hurdles, though evolving rules are beginning to make it easier to use advanced models in pricing.
News: Cyber risks grow more complex for insurers amid AI, human error, and global tensions
The brief: According to Lockton’s 2026 Cyber Threat Report, cyber risk is becoming more complex as AI accelerates attack speed and sophistication, while geopolitical tensions and human error increase exposure. Non‑malicious incidents now represent a growing share of losses, and the biggest financial impacts stem from disruption and third‑party risk rather than ransom. At the same time, rising litigation and privacy claims are exposing gaps in cyber insurance coverage, prompting insurers to tighten underwriting and reassess policy scope.
News: June storms rank among 2026’s costliest insured stretches
The brief: Gallagher Re estimates that severe convective storms across the US from June 1 to 17 will cost insurers in the mid-single-digit billions, with total economic losses running 20 to 25 percent higher once uninsured and underinsured damage is counted. The two-and-a-half-week run of tornadoes, large hail, damaging winds, and flash flooding hit major metros including Chicago, Denver, Dallas, Milwaukee, Washington, and New York City. Gallagher Re called it among the most expensive periods for the industry, regardless of peril, so far in 2026. US severe convective storm losses for the year stood at roughly $22 billion as of June 18, the 11th straight year above $20 billion, though first-half totals remain below the five-year and 10-year averages.
News: AI lowering the barrier to health insurance fraud
The brief: Generative AI is stripping away the skill once needed to commit health insurance fraud, letting bad actors fabricate medical records and run thousands of automated calls a day, an Insurance Business report finds. The National Health Care Anti-Fraud Association estimates up to $480 billion is lost to healthcare fraud each year, and AI already surfaced in last year’s record Department of Justice takedown, which charged 324 defendants over more than $14.6 billion in intended losses. Carriers are answering with their own AI, including voice authentication that flags synthetic callers and tools that detect manipulated medical images down to the pixel. Federal enforcers, meanwhile, are shifting from a “pay and chase” model toward flagging suspicious billing before payments go out.
News: Data centers rising in high climate-risk zones
The brief: A new analysis from climate risk firm XDI finds that a significant number of the data centers being built to support AI are in areas exposed to extreme-weather risk worsened by climate change. Of roughly 2,600 planned data centers worldwide, the study identifies more than 150 on high-risk properties. The build-out is concentrating critical AI infrastructure in places where physical hazards are rising.
News: The widening US flood insurance protection gap
The brief: Moody’s warns the gap between flood risk and flood insurance in the US is wide and getting wider, with the consequences set to worsen for homeowners and state and local governments. The rating agency points to heavier rainfall, stronger storm surge, continued building in flood zones, cutbacks in federal aid, and aging river levees as forces widening the shortfall. As losses climb and coverage lags, more of the cost shifts onto property owners and the public balance sheets behind them.
News: AI joins climate and regulation as top insurer risks
The brief: The eighth Risk Radar from Global Insurance Law Connect, drawing on 26 law firms across 28 jurisdictions, finds AI, regulatory pressure, and climate-related losses converging into a more demanding environment for insurers in 2026. The report points to global insured catastrophe losses of roughly $107 to $108 billion in 2025, with secondary perils including wildfires and severe convective storms dominating the loss landscape. Contributors flag the EU AI Act classifying life and health risk-assessment AI as high-risk from August 2026, while UK and Irish firms describe AI-enabled threats from deepfakes to model manipulation as outpacing policy wordings. GILC chair Gillian Davidson said 2026 will be defined as much by regulators insisting on practical implementation as by the new rules themselves.
News: Allianz tops global AI rankings as agentic adoption surges across insurers
The brief: The 2026 Evident AI Index, covering 30 of the world's largest carriers, finds Allianz overtook AXA for the top global ranking, with Zurich the biggest mover, climbing from 12th to 4th by deploying a shared platform model rather than scattered pilots. Agentic AI surged to one in four newly disclosed use cases, up from one in twenty just six months prior. Three insurers globally now disclose a quantified AI return on investment at group level, and carriers who have made that shift are pulling ahead of those still running point solutions.
News: US government ordered Anthropic to suspend its two most powerful models
The brief: The US government on June 12 ordered Anthropic to immediately suspend Claude Fable 5 and Mythos 5 worldwide, citing national security concerns over a claimed jailbreak. Anthropic complied but publicly disputed the finding, noting the capability in question is already available in other public models including OpenAI's GPT-5.5. The specific concern, as Anthropic describes it, was a jailbreak that prompted the model to read a codebase and identify software flaws, a capability Anthropic says is standard practice in cybersecurity.
News: AI models built for general use are falling short for insurance underwriting
The brief: A Gallagher Re report finds that the AI models insurers are working with were designed for broad general use, not the specialized judgment that underwriting requires. That mismatch makes it hard to accurately price AI-related risks. The cyber side of this is moving fast too, with the time between a vulnerability being discovered and exploited dropping from over two years in 2018 to roughly 10 hours today.
News: States are pushing ahead on AI rules despite White House pressure
The brief: Six months after President Trump warned states not to regulate AI, many are doing so anyway. Congress has still not passed federal AI legislation. California is advancing the "No Robo Bosses Act," which would bar employers from relying solely on AI to fire or discipline workers, while Florida and Utah both stalled under White House pressure. The development-versus-use distinction, where states can regulate how AI is used but not how it is built, is the line to watch for insurance compliance teams.
News: 2026 catastrophe season may outpace carriers' current response models
The brief: Sedgwick's 2026 Catastrophe Season Playbook warns the year's risk landscape is more distributed and harder to predict than carriers have historically planned for. Non-hurricane perils grew total modeled risk by 12% in 2025 versus 2024, and a billion-dollar disaster now occurs every 10 days compared to every 82 days in the 1980s. Twenty-five percent of claims adjusters are expected to retire by end of 2027, with high-turnover carriers already seeing operational costs rise 12%. Sedgwick notes AI and automation have yet to face a true stress test at scale during a major catastrophe event.
News: AI is entering the insurance M&A valuation equation
The brief: PwC's mid-year review finds the insurance sector logged $29.6 billion in M&A deal value across 191 transactions from December 2025 through May 2026, with appetite remaining strong for specialty carriers, MGAs, and E&S businesses. The new variable is AI. Investors and private equity are now directly asking whether AI will allow new entrants to undercut incumbent brokers on cost, or whether incumbents can use it to defend margins. PwC says the answer will shape valuations and capital allocation across the sector through the rest of 2026.
News: AI is now embedded at major carriers, and litigation is next
The brief: Writing in Claims Journal, two attorneys document the shift from AI as an experimental tool to embedded infrastructure at State Farm, Allstate, Progressive, Liberty Mutual, Nationwide, and USAA across underwriting, pricing, and claims. AI is moving underwriting from population-level actuarial groupings to individual-risk prediction, catching behavioral patterns traditional methods miss. The litigation implications follow as carriers use AI to deny claims, plaintiffs' attorneys gain the same tools to challenge those decisions in discovery, and AI-driven denials with high appeal-reversal rates may increasingly surface as evidence of bad faith.
News: AM Best warns data center risks are beyond what P/C insurers have covered before
The brief: A new AM Best report finds the rapid expansion of AI data center infrastructure is creating insurance demand that goes beyond what the traditional P/C market has previously handled. Business interruption is flagged as the most consequential exposure, since a partial or total shutdown of one facility can ripple through interconnected networks. The US has 4,287 data centers as of May 2026, and more than half of planned projects worth $670 billion are in states with high exposure to severe convective storms, tornadoes, and hail. AM Best also identified gaps in builders' risk, environmental liability, cyber, and off-premises power failure coverage as areas requiring new product development.
News: Auto shoppers turn to AI and switch more often
The brief: The J.D. Power 2026 US Auto Insurance Study finds the market has moved from a pricing crisis to an experience problem as rates stabilize. About 32 percent of shoppers now use AI tools when comparing coverage, and those who do are more likely to switch insurers, though 33 percent found the AI content unhelpful. Customer satisfaction is held back by disjointed service across channels. Among the 21 percent forced to cross channels for a single inquiry, satisfaction and renewal intent both drop, and website resolution lags at 66 percent compared with 91 percent once an agent engages.
News: Cyber, AI, and the economy top carrier worries
The brief: RiskScan 2026, a cross-market study from the Insurance Information Institute and Munich Re US covering more than 1,700 participants across five US and UK market segments, ranks cyber incidents, economic pressure, and AI as the top concerns industry-wide. AI was named the most impactful emerging technology, reflecting both fast adoption and rising operational, regulatory, and liability questions. The study flags persistent flood and cyber protection gaps among businesses and professionals alike. It also points to growing recognition of legal system abuse as a driver of rising property-casualty costs.
News: Conference panels flag AI growing pains for insurers
The brief: Panelists at Insurtech Insights cautioned that AI can be overly confident in its output and urged carriers to keep a human in the loop who applies judgment before acting on AI recommendations. Buy versus build has become a key consideration as generative AI lets carriers and brokers solve operational issues themselves, though panelists said insurtechs retain value by focusing on specific operational needs in a consultative partner role. One panel noted insurance carries unlimited downside risk, with loss ratios that can reach 200 or 300 percent, which forces technology leaders to weigh tradeoffs carefully.
News: Moody's says AI will reshape broker profitability
The brief: Moody's Ratings expects AI and technology investment to become increasingly important drivers of productivity and profitability for insurance brokers, with firms that have strong data foundations best positioned to deploy it. The agency sees AI automating policy submissions, reviews, and billing while helping brokers identify new business and assess risk. Moody's cautioned that AI also introduces operational and governance risks including cyber threats, data issues, and model errors, and that gains may not lift profits immediately because firms reinvest in technology, data, and talent. The outlook arrives as organic broker growth slows to low-to-mid single digits in 2026.
News: Draft federal bill would preempt state AI rules
The brief: Democrat Lori Trahan and Republican Jay Obernolte released draft legislation that would bar states from passing laws targeting AI model development, including requirements that models undergo testing before release. The draft would not stop states from regulating how AI technology is used. The pair released the text to gather input from stakeholders and the public before formal introduction. Tech industry group ITI praised the bill, while consumer advocate Public Citizen said it leaves oversight to a federal government that has repeatedly failed to pass meaningful AI protections. The development-versus-use distinction is the line to watch for anyone tracking state AI rules that touch insurance.
News: Former XL chief says insurers must seize AI
The brief: Mike McGavick, the former CEO of XL Group and Safeco, told actuaries at the Casualty Actuarial Society Seminar on Reinsurance that AI offers a major chance to refocus the industry on the problems it exists to solve, but that insurers are falling short today. He framed the opportunity as both an unusual capability and an obligation for carriers to play their role well. McGavick was enthusiastic about the part actuaries can play at the center of that change, particularly in building trust in AI models.
News: Trump signs voluntary frontier AI review order
The brief: President Trump signed an executive order on June 2 directing federal agencies to assess the cybersecurity capabilities of frontier AI models, with AI developers participating voluntarily and the government given up to 30 days of pre-release access. The order creates a clearinghouse where AI companies and agencies, including Treasury and the Pentagon, can exchange information about potential vulnerabilities, and assigns the National Security Agency authority over which models merit early review. Trump had postponed the order in late May over concerns the original 90-day window could undercut US competition with China, and the signed version shortens that window to 30 days. The directive followed Anthropic's same-day announcement that it would broaden access to its Mythos cybersecurity model from a handful of large technology and Wall Street firms to 150 additional organizations.
News: Quiet 2026 forecast still carries severity risk
The brief: Allianz Commercial's 2026 hurricane outlook compiled forecasts from six major meteorological institutions and projects a near to below-average Atlantic season, with NOAA assigning a 55 percent probability of below-normal activity. The outlook warns that storm counts are a poor proxy for losses. In 2025, only 5 Atlantic hurricanes formed, but 4 reached Category 4 or 5, and Hurricane Melissa caused about $11 billion in economic losses across the Caribbean against $2.5 billion in insured losses. Atlantic storms are also intensifying faster, with mean maximum intensification rates roughly 29 percent higher in 2001 to 2020 than in 1971 to 1990, compressing the response window for catastrophe planning.
News: EU appoints expert bodies for AI Act
The brief: The European Commission appointed a 60-member Scientific Panel and an Advisory Forum to provide independent expertise for enforcing the EU AI Act. The Scientific Panel brings together experts in frontier AI, engineering, technical auditing, industry, and societal impact, and will focus on general-purpose AI models and systems, systemic risks, model classification, evaluation methodologies, and cross-border market surveillance. The Advisory Forum will provide independent technical expertise on a broad range of AI Act issues including standardization and implementation challenges, drawing members from academia, civil society, industry, startups, and SMEs. Members of both bodies will serve two-year terms.
News: Insurers unprepared for AI-driven fraud
The brief: Data and AI firm SAS warned that generative AI tools now let virtually anyone with a computer create or alter images to file fraudulent insurance claims, from fake crash scenes and damaged furniture to altered receipts. About 1 in 10 property-casualty insurance losses already involves fraud, and SAS said growing access to image-generation tools will push that figure higher by lowering the technical skill needed to fabricate convincing evidence. A joint survey from the Association of Certified Fraud Examiners and SAS found only 7 percent of anti-fraud professionals say their organization is more than moderately prepared to detect AI-driven fraud, and no insurance respondent expressed more than moderate confidence. SAS noted AI also offers a path forward for insurers, since it can analyze huge volumes of claims data and detect image anomalies that humans cannot.
News: AI fueling new wave of malpractice claims
The brief: AI has moved from a theoretical concern to an active source of legal malpractice claims, according to EPIC Insurance Brokers and Consultants' 16th Annual Lawyers' Professional Liability Claims Survey, which polled senior claims executives at 13 lawyers' professional liability insurers that together cover more than 80 percent of firms in the Am Law 200. Seven of the 13 insurers reported an increase in AI-related claims during the past year. Eight of the 13 insurers reported higher claim frequency compared with the previous year, and respondents cited continued growth in large-loss claims, including matters reserved or paid in excess of $100 million. A majority of surveyed insurers indicated plans to raise rates in 2026, with 11 of 13 reporting materially higher defense spending year over year.