The key AI and insurance stories driving market activity, regulatory developments, and operational change this August.
Latest Articles as of August 27
News: Cyber insurers clarify how policies cover autonomous AI
The brief: Cyber insurers are reviewing and updating policy language as autonomous AI agents raise new questions about what counts as a covered event. Insurers are largely clarifying how existing coverage applies rather than adding exclusions. QBE's global head of cyber said the insurer treats AI as a risk amplifier rather than a fundamentally new category, and that AI-related events leading to a conventional cyber incident stay within a cyber policy. Targeted AI cover already exists from providers like Armilla, Munich Re's AiSure, and AXA XL for risks such as model underperformance and hallucinations. The global cyber market, worth nearly $15 billion last year, is expected to reach roughly $28 billion by 2030, per Munich Re.
News: Claims leaders weigh AI's role in both driving and catching fraud
The brief: AI-enabled fraud and data authenticity is the factor most likely to increase claim severity over the next three to five years, named by 32% of 2026 RIMS conference attendees. The same survey found governance concerns, including privacy, security, compliance, and model risk, are the biggest barrier to using AI effectively in claims, cited by 33% of respondents. CorVel noted that responsibly implemented AI can also work the other way, helping claims teams spot the fraud patterns and anomalies that would otherwise slip through. It pointed to layered verification, continuous monitoring, and strong governance as the practices that make that work.
News: Legal & General presses Meta for clearer AI data-center disclosure
The brief: Legal & General's asset-management arm, which oversees roughly £1.2 trillion, disclosed that it backed shareholder resolutions this year asking Alphabet, Amazon, and Meta to explain how their AI infrastructure ambitions square with their stated emissions-reduction goals. The resolutions did not pass, but L&G singled out Meta, saying that against its peers the company could improve disclosure on how it balances speed-to-power against clean and reliable power. L&G also sought more clarity on the role of natural gas in Meta's decarbonization pathway. Meta pointed to its existing environmental reporting and said the requested report would not add benefit for shareholders.
News: Agency buyers start pricing AI readiness into deals
The brief: Insurance brokerage buyers are beginning to factor a target's technology and AI capability into how they assess acquisitions, even as valuations stay anchored to growth. MarshBerry's 2026 Technology and Governance Report found 82% of respondents expect AI to affect broker productivity or reshape the broker's role, while only 13% actively use AI tools across multiple parts of the business. Reagan Consulting data showed 11.5% of firms under $1.25 million in revenue invested in AI in 2025, against 84.2% of brokers above $100 million. One executive estimated broad AI implementation could lift a person's capacity by 50% to 75% on tasks like submissions and proposal prep, with the plan to redeploy that capacity rather than cut headcount.
News: Baker Tilly reframes vendor AI oversight as an 'express lane'
The brief: John Romano, a Baker Tilly principal who leads its insurance regulatory practice, argues carriers should replace a one-size "toll booth" review with an "express lane" that defines AI risk appetite up front, letting low-risk vendor uses move quickly while high-consumer-impact uses draw deeper scrutiny. He tiers vendors by consumer impact, placing MGAs with delegated authority at the top, followed by TPAs and claims-technology providers. Romano starts any review with an inventory of where a vendor uses AI, what decisions it makes, and where humans stay in the loop, rather than reading a stated policy. He expects regulators to move from documentation toward evidence of the specific controls that matter most in underwriting, pricing, and claims.
News: Munich Re buys cyber insurtech At-Bay for $575M
The brief: Munich Re Group agreed on August 19 to acquire cyber insurtech At-Bay at an enterprise value of $575 million, with closing expected in the first quarter of 2027 after regulatory approvals. At-Bay, which primarily serves small and mid-sized US enterprises and provides managed detection and response alongside coverage, will be overseen by Hartford Steam Boiler, part of Munich Re's Global Specialty Insurance business and a partner since At-Bay's 2017 founding. Munich Re board member Mike Kerner called At-Bay a strong fit for its specialty portfolio and expects it to become an earnings growth driver over time. Munich Re framed the deal as positioning the group for a cyber market moving from standalone coverage toward integrated, continuously managed risk platforms.
Latest Articles as of August 20
News: AI is helping cyber defenders move faster than attackers
The brief: Beazley Security's Q2 2026 Quarterly Threat Report found that agentic AI used in vulnerability research drove a 36% quarter-over-quarter jump in newly disclosed vulnerabilities, while vulnerabilities confirmed as actively exploited rose just 10%. Beazley reads this as AI currently benefiting research faster than it aids attackers. CEO, Alton Kizziah, said attackers are experimenting with AI but still lean on tried-and-true methods, with compromised credentials behind 67% of ransomware intrusions Beazley investigated. Its advice is that strong credential hygiene and assessments of how AI is used inside an enterprise remain the highest-value moves.
News: AI is helping cyber defenders move faster than attackers
The brief: Beazley Security's Q2 2026 Quarterly Threat Report found that agentic AI used in vulnerability research drove a 36% quarter-over-quarter jump in newly disclosed vulnerabilities, while vulnerabilities confirmed as actively exploited rose just 10%. Beazley reads this as AI currently benefiting research faster than it aids attackers. CEO, Alton Kizziah, said attackers are experimenting with AI but still lean on tried-and-true methods, with compromised credentials behind 67% of ransomware intrusions Beazley investigated. Its advice is that strong credential hygiene and assessments of how AI is used inside an enterprise remain the highest-value moves.
News: AI could refresh catastrophe models in days, not months
The brief: Karen Clark & Company says AI could change how catastrophe models are built and updated, giving re/insurers more current risk insight for pricing, underwriting, claims and capital decisions. In an August white paper, KCC describes a shift toward high-resolution physical models enhanced by AI, especially for frequent perils like severe convective storms and wildfires where historical data falls short. Its Severe Convective Storm model already processes more than 30 gigabytes of satellite, radar, and weather data a day. KCC says the biggest benefit is speed, with models updating in days rather than months so carriers can price off fresher information.
News: Gallagher Re builds a unit to close the AI coverage gap
The brief: Gallagher Re has launched a digital risk practice that pulls AI liability, data centers, cyber, and digital risk engineering into one advisory function, responding to demand as technology exposures increasingly cut across multiple lines at once. The firm points to the July 2024 CrowdStrike outage, which caused roughly $5.4 billion in economic damage to Fortune 500 companies against far smaller insured losses, as the kind of gap the practice is designed to close. Demand is real, with a Munich Re survey finding 63% of C-level executives want to buy insurance against AI-related risks. Gallagher Re frames this as translating new exposures into underwriting insight rather than backing away from them.
News: AI adoption could drive a wave of insurance dealmaking
The brief: PwC's Mark Friedman told InsuranceNewsNet that the ability to adopt and scale AI could be a major driver of insurance M&A over the next 12 months, as companies that lack the scale to invest look to combine or get acquired. Friedman argues AI will expand what insurers can cover, not shrink it, predicting new products and more insurable risk as AI reshapes every sector. He also sees an agentic workforce running back-office operations as a path to lower costs and higher profitability for carriers that move early.
News: AI is becoming small businesses' first stop for coverage answers
The brief: A new ERGO NEXT Insurance survey of 501 US small business owners found that 36% have used an AI tool to research coverage, nearly matching the 31% who consulted a human agent. Among those who used AI for insurance guidance, 94% said they trust the information they received. The survey frames AI as additive to agent relationships rather than a replacement, with owners using AI as a starting point while still turning to agents for complex decisions. It also surfaces a real opening, with 73% of small businesses carrying inadequate coverage that AI plus human expertise could help close.
News: Carriers build the tools to keep writing AI risk
The brief: Three new ISO endorsements for generative AI (CG 40 47, CG 40 48, and CG 35 08) have drawn growing carrier interest, giving underwriters clearer options for handling AI exposure in commercial general liability. Verisk's Joe Lam, who helped write the endorsements, frames them as underwriting flexibility that lets carriers accept a risk for the right premium rather than walk away from it, which he expects to add stability to the market. Attorney Alana McMullin notes some insurers may choose to price and underwrite AI risk for additional premium instead of excluding it.
Latest Articles as of August 13
News: Only 6 percent of insurers qualify as AI leaders
The brief: A new EXL U.S. Enterprise AI Study of more than 300 respondents finds a wide gap between how insurers rate their AI progress and where they actually are, with 76 percent believing they lead the competition even though only 6 percent qualify as leaders. Scaling AI is now a priority for 96 percent of insurers, up from 86 percent in 2025, and agentic AI is advancing fastest in risk management, actuarial, underwriting, and customer experience. Leaders generated 40 percent more revenue growth and 37 percent more cost reduction in use cases where AI was applied. Data is the top obstacle, cited by 92 percent of insurers, with only 24 percent rating themselves mature on data management.
News: Underwriters reassessing coverage for autonomous cyber attacks
The brief: An Israeli cyberdefense firm says a hacking tool built from open-source AI agent frameworks ran largely without human input inside Taiwanese government systems over four days in early July, first reported by the Financial Times. Up to eight agents worked at once, mapping 21 systems, compromising at least 85 accounts, taking more than 2,500 personnel records, and expanding into a nuclear safety agency and seven energy companies. Some policy forms still define a hacker as a person, which Browne Jacobson's Tim Johnson notes leaves open whether an autonomous AI attacker triggers coverage at all, and the murky attribution complicates nation-state exclusions. Brokers at Assured and Brown & Brown expect carriers to revisit widespread-event exclusion language as autonomous tools change how a single attack can scale.
News: AI buildout driving record demand across P&C
The brief: AIG CEO said that the AI data center buildout is a "great opportunity" for insurers but is also pushing property and casualty capacity to its limits. He said these projects need coverage across project finance, construction, cyber, property, and liability, and that AIG has a competitive advantage because it offers products in all of these areas. AIG CEO said the required limits are absolutely maxing out the P&C industry. AIG is involved in several parts of the construction process and also provides insurance to power providers. AIG is also looking at AI to streamline its own underwriting and claims work.
News: Satellite links leaving superyachts cyber exposed
The brief: Cheap low-earth-orbit satellite connectivity has brought superyachts online in a way older VSAT systems never did, and cyber resilience has not kept pace, according to Intersys managing director, Matthew Geyman. Navigation and propulsion systems built on protocols that predate always-on connectivity are now reachable by basic network scanning, while high crew turnover and weak password hygiene leave controls thin. Geyman says resilience now matters more than insurance, since good controls reduce both financial loss and physical risk to crew and guests, and he points to the long-standing CL380 clause that excludes cyber from hull and P&I cover. Dedicated marine cyber products have emerged, but underwriters are scrutinizing vessel resilience more closely.
News: Cyber underwriters starting to ask about client AI use
The brief: UK cyber insurers are aware their clients' staff are feeding sensitive data into public large language models, but the market has not agreed how to underwrite it, according to a panel of cyber specialists speaking to Insurance Business TV. Jensten London Markets', Daniel Winn, said some carriers now offer full affirmative AI cover up to policy limits while others cap it, with one or two setting the ceiling around £250,000, reflecting a shortage of claims data rather than a settled view of the risk. A Talbot Jones broking manager said there is no consistent set of underwriter questions on how clients actually use AI, leaving brokers to volunteer what they judge material rather than answer anything formally asked.
News: AI firms captured 99% of Q2 InsurTech funding
The brief: Global InsurTech funding reached $2.44 billion in the second quarter of 2026, the highest quarterly total since Q2 2022, with AI-focused companies capturing 99.1 percent of it across 95 deals, according to Gallagher Re's latest Global InsurTech Report. Every round above $5 million went to an AI-focused company, though the money concentrated in fewer businesses and was driven more by venture and private equity capital than by insurer-backed investment. The same report says the AI data center buildout as one of the largest new pools of insurable assets in decades, spanning construction, chip transport, and business interruption from power and cooling failures.
Latest Articles as of August 6
News: Moody's sees AI benefits arriving gradually, with new risks
The brief: Moody's Ratings expects AI to lift insurer productivity and cut operating costs over time, but frames the gains as material and gradual rather than transformative. The firm sees the biggest near-term disruption in retail P&C distribution, where high transaction volumes and standardized products suit automation, while life insurers face a more limited impact given product complexity and longer liabilities. Moody's says that competitive markets may force insurers to pass efficiency savings to customers through lower prices, and that AI raises operational, regulatory, litigation, and cybersecurity risks tied to model opacity, algorithmic bias, and dependence on a small set of third-party providers. It expects performance to diverge, favoring insurers with strong data infrastructure and the ability to redesign processes around AI.
News: Insurance AI pilots stall when no one owns them after launch
The brief: Forestview Insights' CEO told Insurance Business that the hardest part of insurance AI is no longer proving a model works, but deciding who owns it once the pilot ends. He describes a common failure mode where a project succeeds, the data scientists and business experts move on, and no one is left to handle monitoring, retraining, patches, or support. Launch budgets often cover development and implementation but not the ongoing lifecycle costs, even though data, workflows, and regulatory expectations keep shifting. The piece cites BCG's finding that only 7% of insurers had scaled AI successfully as of 2025, and argues the carriers that win will convert pilots into maintained, funded, business-owned products.
News: Rights groups and state AGs push back on FTC's AI policy
The brief: A coalition including the Electronic Frontier Foundation, the Center for Democracy and Technology, and attorneys general from California, Colorado, and roughly 20 other states filed formal comments on August 3 urging the FTC to withdraw its proposed policy statement on AI accuracy. The FTC's proposal, published in July, argues that AI companies steering their systems' outputs away from what consumers expect could be committing deception under Section 5 of the FTC Act. Critics counter that the proposal rests on a false premise, since no neutral baseline output exists, and warn it could expose routine safety tuning to federal liability. The state attorneys general argue the approach would force companies to choose between conflicting state and federal obligations. For insurers building AI governance around a patchwork of state rules, the dispute underscores how unsettled the federal-versus-state picture remains.
News: Underwriters now weigh an employer's AI strategy when choosing jobs
The brief: A Sixfold survey of 543 underwriting executives and underwriters across the US and Europe found that 72% would weigh an employer's AI strategy when considering a new role, and 69% said their company's approach to AI makes them more likely to stay. The report, covered by Insurance Journal, frames AI strategy as a retention issue, with 77% of executives worried about losing underwriters to competitors with stronger AI tools. Nearly all executives and 94% of underwriters said AI has made their teams faster, better, or both, though 83% of underwriters still spend at least 30 minutes chasing missing information from brokers on a single risk. Notably, not a single underwriter said they treat AI as a fully trusted part of how they work.
News: Carriers must decide whether AI expands access or just segments risk
The brief: Birny Birnbaum, of the Center for Economic Justice, told Insurance Business that the most important AI question for insurers is what they are optimizing for, not what a model can predict. He says that AI holds real promise for making insurance more available, more affordable, and faster to service, but warns the same capability can be used to further segment a book, avoid certain risks, and leave consumers with fewer affordable options. Birnbaum points to climate risk, where insurers have often responded by cutting coverage and shifting risk onto consumers and public programs, and argues AI can either reinforce or challenge that pattern.
News: Brown & Brown says AI push won't drive major layoffs
The brief: Brown & Brown CEO, Powell Brown, said the brokerage's "AI-first" rollout across roughly 23,000 employees is a growth tool, not a cost-cutting exercise, and does not anticipate significant AI-related workforce reductions. Brown expects roles to change and employees to reskill as the technology embeds, with early use cases including dataset analysis, knowledge sharing, and automating repetitive work such as certificates of insurance. Brown said senior-leadership commitment and treating the effort as a business transformation rather than an IT project are central to adoption.
