Check out the latest news and insights to stay in the know about the trends and innovations driving the insurance industry forward.
Latest Articles as of July 30
News: Germany moves to oversee how insurers use AI
The brief: Germany's financial regulator, BaFin, gained authority this week to oversee how banks and insurers use AI, under national legislation implementing the EU AI Act that took effect July 29. Starting immediately, BaFin will monitor compliance with transparency rules, including whether firms disclose when customers are dealing with a chatbot, and watch for prohibited practices such as analyzing sensitive personal data in ways that unfairly disadvantage people. Tighter oversight of higher-risk systems, including those insurers use for risk assessment and to price life and health policies, does not begin until December 2027. BaFin said it will review a sample of AI applications rather than every system, and can levy fines up to €35 million or seven percent of turnover in extreme cases.
News: Machine learning may be quietly breaking discrimination law
The brief: Underwriting algorithms find and exploit correlations with race, gender, and other protected characteristics even when those inputs are excluded, a University of Minnesota law professor told Insurance Business, describing proxy discrimination as the highest-order liability in AI governance today. Denying the AI intuitive proxies just pushes it toward less intuitive ones, he said, and competitive pressure to price more precisely makes the practice hard to resist. The NAIC is piloting an AI Systems Evaluation Tool across 12 states to standardize how examiners review insurer AI programs. The reporting frames the regulatory direction as moving toward documented testing, not just written policies.
News: Data breach costs hit a record as AI reshapes cyber risk
The brief: IBM's annual Cost of a Data Breach Report puts the global average breach at nearly $5 million, up 12 percent, with the US average at a record $11.5 million. AI-driven attacks rose 56 percent over last year's study and added roughly $1 million per breach. IBM found that 92 percent of organizations hit by an AI-related breach lacked proper AI-access controls, though 85 percent said they plan to raise spending on security tools and governance. Three quarters said they will lean harder on AI agents for alert triage, vulnerability management, and penetration testing. For cyber underwriters, the report shows AI as both a cost driver and a control that insureds are increasingly expected to have in place.
News: 7 states restrict AI in health coverage decisions
The brief: Seven states enacted laws this year restricting how health insurers use AI in coverage and reimbursement decisions, according to a Becker's Payer roundup. The common thread is that AI cannot be the sole basis for a medical-necessity denial, with Alabama, Georgia, Iowa, Utah, and Washington all requiring human review of adverse determinations and disclosure when AI is used. Colorado's HB 1139 goes further, requiring utilization-review AI to weigh a patient's individual clinical history rather than group data, to be audited for accuracy, and barring insurers from covering AI-delivered psychotherapy. Illinois added a Transparency in Downcoding Act that blocks automated processes from bypassing a billing professional's submitted codes. Effective dates run from July 2026 through January 2028.
News: The Hartford says AI is compressing underwriting time
The brief: The Hartford is deploying AI-enabled underwriting across middle-market and large commercial lines, with the CEO telling analysts on the carrier's Q2 earnings call that underwriting activities are being completed in a fraction of the time. Swift said the gains lift productivity and free underwriters to spend more time building agent and broker relationships to drive submission flow, while stressing that underwriters continue to own the decision. The company also applies the capabilities in claims and customer-facing operations. The Hartford reported second-quarter net income of $1.3 billion, up 30.5 percent year over year, on revenue of $7.26 billion.
News: AI could reset insurance's two decades of flat growth
The brief: A McKinsey report argues AI could break the structural stagnation that has defined insurance since 2005, when premiums grew about 4.9 percent a year but profits trailed at 4.3 percent. The firm points to new risk pools AI is expanding faster than current products can serve, including AI liability, nonphysical business interruption, and cyber, where McKinsey notes less than 1 percent of global cyber costs are currently insured. It also flags a distribution shift, with agentic AI increasingly acting as the customer's front door and shopping renewals across carriers before a person gets involved. McKinsey frames precise pricing and claims accuracy on volatile risks as the capability that lets carriers write coverage others decline.
Latest Articles as of July 23
News: Berkley reports 20%+ underwriting efficiency from AI
The brief: W. R. Berkley CEO, Rob Berkley, told the company's Q2 2026 earnings call that AI is delivering efficiency gains of more than 20% in underwriting, with work focused on underwriting workbenches and digitizing the path from intake to quote. He said there is more room to run and pointed to claims as the next target, noting that roughly half of Berkley's claims settle for $5,000 or less and that AI can route those simpler files toward straight-through processing. Berkley was direct about strategy, saying the firm will not build its own large language model but will layer its own approach on top of existing tools and test them across its many operating units. The comments put hard numbers to a carrier-scale deployment rather than a pilot.
News: Carriers move to write AI out of standard liability forms
The brief: Insurers are increasingly attaching generative-AI exclusions to commercial liability policies, with three ISO exclusion forms drawing rising carrier interest, according to a cyber-focused attorney and an expert who helped draft them. Lathrop GPM partner Alana McMullin said the industry has shifted sharply toward limiting AI exposure as AI-related lawsuits become more frequent. The exclusions attach to commercial general liability coverage and address losses tied to policyholders' use of the technology. The move points to a widening gap between how quickly businesses are adopting AI and how willing carriers are to cover the fallout.
News: An underwriter's take on AI cutting hours to seconds
The brief: A personal lines underwriter and broker at SPG Wholesale, said AI has changed excess and surplus lines work less by automating decisions than by speeding up how fast he can check whether a risk fits a carrier's guidelines. Submission volume has climbed sharply, driven partly by West Coast wildfire activity pushing more business into the E&S market, and he said AI tools now let him scan hundreds of pages of carrier guidelines in seconds rather than hours to reach a fit-or-no-fit answer. He stressed that client relationships still depend on human connection AI cannot replicate. Looking ahead, he said communication skills paired with fluency in AI tools will set apart underwriters entering the field over the next five years.
News: Global catastrophe losses ran well below average
The brief: Aon estimated global insured catastrophe losses at $47 billion for the first half of 2026, down sharply from $108 billion a year earlier, with severe convective storms again the costliest peril at roughly $27 billion in the US. US natural catastrophes drove about 75% of global insured losses, and the country logged 11 billion-dollar events, nine of them convective-storm outbreaks. A late-April US outbreak was the largest single event of the half at more than $5 billion, and both Illinois and Indiana set annual tornado records by midyear. The below-average total reflects lower storm frequency and a more fragmented loss pattern rather than any easing of exposure.
News: A viral AI spat turns the lens on claims denials
The brief: A public exchange between Mark Cuban and Marc Andreessen over whether AI has surpassed human doctors drew fresh attention to how health insurers use AI on the claims side. NAIC survey data cited in the piece found 84% of health insurers use AI or machine learning in some capacity, 12% of them to deny prior authorizations, and nearly a third do not regularly test their models for bias despite NAIC guidance. The governance gap is showing up in litigation, with Cigna, UnitedHealthcare, and Humana all facing suits over algorithmic claims handling, and Cigna's case allowed to proceed in March 2025. California already bars AI-only denials of medically necessary care, and the NAIC's evaluation tool is piloting in 12 states ahead of a national rollout.
News: Cyber outlook holds steady as AI reshapes the risk
The brief: AM Best kept its stable outlook on the global cyber insurance segment, citing continued demand despite softer pricing, favorable medium-term profitability, and the growing role of AI in the market. The agency expects underwriting margins to stay healthy even as competition narrows them, supported by better policyholder cyber hygiene and ample reinsurance and insurance-linked securities capacity. AM Best flagged persistent headwinds, including ransomware, business email compromise, and funds-transfer fraud, all of which it says are escalating and driving large-loss potential. It also cited the February 2026 Iran-US conflict as a likely demand driver, while noting that standard war exclusions may limit some related exposure.
Latest Articles as of July 16
News: UK elevates AI cyber and election-interference threats officially
The brief: The UK added AI-driven cyber attacks and a new "digital resilience failure" risk to its National Risk Register on July 15, drawing on lessons from the 2024 CrowdStrike IT outage. Munich Re values the UK cyber insurance market at $1.75 billion in 2025 and forecasts continued growth as the Cyber Security and Resilience Bill expands obligations to critical suppliers. The National Cyber Security Centre logged 204 significant cyber incidents through September 2025, more than double 2024's 89, with attacks on Marks & Spencer, the Co-op, and Jaguar Land Rover producing business interruption losses. Chief secretary to the Prime Minister Darren Jones told Parliament that AI creates new criminal cyber attack vectors alongside opportunities for the economy.
News: Plaintiffs' bar outpaces defense on AI
The brief: Plaintiffs' firms are using AI to identify high-value claims, coordinate strategy across firms, and generate large volumes of discovery requests that pressure insurers before investigations finish, according to defense attorneys and claims experts speaking to Insurance Business. Tyson & Mendes senior partner Ashley Fetyko said the plaintiffs' bar adapts quickly while insurance defense is slower, with recognizable AI-generated language now surfacing across unrelated cases. Sedgwick's Summer 2026 Liability Litigation Observations and Trends report clocks the average first-notice-of-loss-to-filing time at roughly 120 days, down from 550 days in 2016, with 70 percent of bodily injury claimants retaining counsel within two weeks. Third-party litigation funding in auto liability has grown 44 percent annually since 2022.
News: AI exposure hides in unpriced policy language
The brief: More than 90 percent of insurers' AI agent exposure sits inside conventional policies never designed for the technology, concentrated in cyber, D&O, commercial general liability, and technology errors and omissions lines, according to Underwriting the Agent Economy, a new report from the Artificial Intelligence Underwriting Company. Willis research cited in the report found professional liability carriers shifted between January 2025 and January 2026 from silent AI assumptions to either affirmative warranties or absolute exclusions. AI agents differ from chatbots because they carry out tasks rather than generate responses, operating software, accessing company data, and moving funds with limited human oversight. Aon head of intangible assets Kevin Kalinich, said AI could produce "aggregated, systemic, correlated" losses.
News: Loss prevention emerges as AI's underwriting payoff
The brief: Insurance executives at a Lloyd's Lab roundtable said AI's greatest value in property and casualty so far is changing how underwriters evaluate risk to help customers avoid losses before they occur. Dawn Miller, CEO of Lloyd's Americas and chief commercial officer at Lloyd's, said enriched data lets underwriters make more creative decisions, close protection gaps, and take on risks that could not be handled before. Rohit Bhat, managing director of financial services at Google Cloud, said the emphasis is moving from administrative automation and customer interactions toward core risk selection and prevention. Panelists framed the shift as a move away from paying out after events and toward reducing their impact in the first place.
News: P&C insurers log strongest year since 2015
The brief: The US property casualty industry posted its best underwriting profit and combined ratio in a decade in 2025, with direct premiums written growing 5 percent to about $1.11 trillion and the industry combined ratio finishing at 93, according to a new AM Best report. Underwriting income jumped from about $23 billion in 2024 to $61.2 billion in 2025 on the strength of personal lines, with personal auto more than doubling its underwriting income. Commercial auto improved from a $4.9 billion underwriting loss in 2024 to a $1.9 billion loss in 2025, with liability losses still mounting and another $2 billion in reserve deficiencies recorded. AM Best's associate director attributed the personal-lines improvement to insurers applying technology and data analytics to underwriting, claims handling, and rate making.
News: EPL market fractures over AI hiring exposure
The brief: Employment practices liability carriers are splitting three ways on AI coverage as Mobley v. Workday advances in federal court, with a June 2026 ruling allowing key California and federal disability discrimination claims to proceed against Workday's AI-powered hiring tools. Absolute AI exclusions now appear in roughly 10 percent of the EPL market, generally written into management liability common terms and conditions where they can also reach D&O and E&O claims, Burns & Wilcox broker Lucas Roberts said. Most policies stay silent and treat algorithmic and human misconduct alike, meaning EPL coverage responds to AI-driven discrimination the same as it would for a hiring manager's, Travelers product manager Chris Williams said. A handful of carriers, including Counterpart, are writing affirmative AI coverage.
News: Buyers gain leverage as reinsurance capital swells
The brief: Property catastrophe reinsurance buyers secured risk-adjusted price reductions of 15 to 25 percent on US treaty placements and 20 to 40 percent on property facultative reinsurance at mid-year renewals, with global reinsurer capital reaching a record $790 billion at March 31, 2026, according to Aon's Reinsurance Market Dynamics Midyear 2026 report. Global reinsurance demand rose more than 10 percent, driven by Florida-based insurers who added an estimated $5 billion to $7 billion of new coverage. Capacity was more than adequate to absorb the growth, with both traditional reinsurers and insurance-linked securities investors deploying capital aggressively. Aon said improvements in data, analytics, and AI are helping expand capacity and support more customized reinsurance solutions across the market cycle.
Latest Articles as of July 9
News: OpenAI ends federally requested access limits on GPT-5.6
The brief: OpenAI said it will publicly release its GPT-5.6 Sol, Terra, and Luna models, ending weeks of restricted access at the request of the US government under the June AI executive order. The order asked developers to voluntarily submit frontier models for federal safety review before wider release, part of the same framework insurance regulators and enterprise buyers are watching as they set AI governance policies. OpenAI said GPT-5.6 Sol is its strongest model to date, with the company citing gains in coding, biology, and cybersecurity. The launch follows a similar path taken by Anthropic, which restored access to its latest models last week after a weeks-long clash with the government.
News: El Niño may shift storm and wildfire risk in unexpected directions
The brief: NOAA formally declared El Niño in June and puts the probability of an extreme event above 60% by winter, with catastrophe modeler Karen Clark and Company publishing a July whitepaper on what that could mean for insured losses. KCC noted El Niño historically reduces North Atlantic hurricane activity but cautioned that the correlation between ENSO phase and insured losses is weak, since the biggest driver of losses is where storms strike, not how many form. The report flagged a possible increase in severe convective storm activity in the US Southeast in spring 2027 if the pattern persists, alongside potentially reduced wildfire activity in California from El Niño-linked atmospheric rivers. KCC estimated a single major hurricane striking Miami, Tampa, or Houston could still produce insured losses exceeding $100 billion at today's exposure levels.
News: Illinois signs first-in-nation AI safety audit law
The brief: Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act on July 6, making Illinois the first state to require mandatory annual third-party safety audits of large AI developers. The law, modeled on California and New York bills but going further on the audit requirement, obligates covered developers to publish risk-mitigation frameworks, report significant safety incidents within 72 hours, and maintain whistleblower protections. Illinois lawmakers estimated California, New York, and Illinois together account for roughly 40% of the US AI market, positioning the trio as a de facto national framework. The law takes effect January 1, 2028.
News: Travelers used AI and geospatial imagery on Palisades and Eaton fire claims
The brief: Travelers' 2025 Sustainability Report described a phase of AI deployment the carrier calls "Innovation 2.0," with more than 20,000 employees now using AI tools regularly and dozens of generative AI applications in production. Following the January 2025 Palisades and Eaton fires, Travelers used high-resolution aerial imagery, geospatial analytics, and AI to identify total-loss properties and route claims to trained disaster specialists. The company said it resolved 90% of property catastrophe claims within 30 days across 2025, despite responding to 62 catastrophe events and over 80,000 catastrophe notices of loss. That performance came against a $2.266 billion first-quarter catastrophe loss, of which $1.7 billion was tied to the California wildfires.
News: Over half of states enact new AI laws
The brief: The NYU Center on Technology Policy reports that as of July 1, more than half of US states have enacted over 100 new AI laws this term, with 109 AI laws and 28 data center laws counted so far. Insurance features directly in the consumer-protection wave. States enacted at least six new laws restricting the use of AI by health insurers, spanning both Republican-led states such as Iowa and Democrat-led states such as Washington. Several states, including Connecticut and Nebraska, also passed laws limiting AI-enabled dynamic pricing. The Center notes that broad algorithmic-discrimination rules have largely collapsed after Colorado scaled back its 2024 AI Act, even as narrower consumer-protection provisions expanded.
News: Four in ten insurers now use AI in underwriting
The brief: A new Sollers Consulting study spanning ten markets reports that 40% of insurers now use AI in underwriting, with the technology reaching a function that historically lagged both management and distribution in digitalization. Sollers identified 126 active AI use cases in the sector, of which 13 focus on underwriting, and noted the sharpest advancements are in commercial insurance, where roughly one-fifth of providers use AI to process data from unstructured documents and triage submissions. Sollers head of underwriting Jakub Śliwiński said extending AI to more complex, lower-volume risks will take another one to two years, since the data foundations still need to be built.
News: Insurance leads all sectors in AI readiness gains
The brief: Accenture's inaugural AI Progress Barometer, which tracks roughly 3,000 of the world's largest companies on a 0-to-100 readiness scale, ranked insurance as the fastest-improving sector globally with an eight-point gain to 48.6 over six months. Ten of the 18 sectors tracked improved overall, with travel (+5.7) and consumer goods (+5.2) trailing insurance. European companies improved 1.6 points versus 1.1 in North America, though North American firms still lead on absolute readiness at 48.9 versus 43.1. Accenture attributed the insurance sector's gain to carriers redesigning workflows around AI rather than layering it on top of existing processes, with clean data and workforce training singled out as prerequisites.
Latest Articles as of July 2
News: Carriers win approval to exclude AI liability
The brief: State regulators across all 50 states have approved carrier requests to exclude AI from standard liability policies, with more than 80% of such requests cleared. Major writers including Berkshire Hathaway, Chubb, and Travelers have won approval to strip AI-caused damages from corporate coverage, and standardized ISO general liability forms now offer a clean generative AI exclusion. Insurers point to the lack of loss history as the reason they cannot yet price the exposure.
News: Cyber insurance loss ratio rises again
The brief: AM Best reports the US cyber insurance loss ratio rose for a second straight year to 53 in 2025, its first reading above 50 since the pandemic ransomware spike. Premium was essentially flat, and the first quarter of 2026 marked the eighth consecutive quarter of pricing cuts, which AM Best says will make the rising loss ratio hard to reverse. Third-party claims are trending up about 30% and carry a longer tail, adding uncertainty to future losses. Surplus lines carriers, now nearly two-thirds of cyber premium, are positioned to absorb much of that development.
News: AI hallucinations in court raise insurer exposure
The brief: A legal analysis says that generative AI hallucinations in court filings are creating fresh exposure for insurers writing lawyers professional liability, E&O, and cyber cover. It points to a recent case where solicitors referred themselves to their regulator after the court was misled by AI-generated citations, with the judge focused on supervision and verification failures. The firm argues courts and regulators are treating AI errors as process failures rather than technology surprises. For insurers, that pushes AI risk management beyond underwriting models into panel counsel oversight, audit rights, and policy wording.
News: Data center build-out raises workers' comp exposure
The brief: The race to build AI data centers is deepening the US skilled-labor shortage, pushing contractors to hire and train less experienced workers faster, which a Zurich executive links to higher workers' compensation exposure. First-year employees consistently show higher injury rates, so carriers are working with contractors and brokers on loss control for the projects that need the most skilled trades. Zurich reports the strain is showing up mainly in scheduling and project delays rather than poorer workmanship. The insurer is also deploying job-site video analysis from construction technology firm Aerosite to flag unsafe behavior during the highest-risk phases.
News: NAIC breach pauses insurer investment designations
The brief The NAIC said an attacker exploited a zero-day vulnerability in Oracle PeopleSoft to reach part of its environment on June 11, as part of a broad campaign hitting many organizations, and has since published the stolen data. Based on findings to date, the exposed data included already-public statutory financial filings and credit rating agency determinations on insurer investments, with no current evidence that personal or financial account information was taken. State insurance department systems and core regulatory filing platforms including SERFF, OPTins, and UCAA were not compromised. Because some credit rating agencies paused their data feeds, the NAIC temporarily suspended assigning designations to insurer investments and told insurers to watch AVS+ for updates.
News: AI pushes tech risk up boardroom agendas
The brief: A Clyde & Co survey of 700 senior decision-makers across 11 industries found 86% now rate technological risk as high impact, up from 46% a year ago, the largest jump of any category. A Clyde & Co partner said AI governance frameworks struggle to keep pace, since a framework that looks mature today can be out of date tomorrow. Nearly three-quarters cited technology implementation and systems integration among their greatest operational challenges, and more than half expect tech adoption including AI to pose significant risk over the next year. Geopolitical risk also climbed sharply, with 72% reporting direct commercial impact versus about half in 2025.
