Blog

How Confidence Scoring and Thresholds Keep Insurance AI in Check

Roots Experts — Wed, 27 May 2026 16:28:28 GMT

Across every functional layer of insurance – underwriting, claims, servicing, and beyond – AI-driven workflows are fast becoming central to how insurance work gets done. But as these systems take on greater responsibility, anyone deploying or overseeing them faces a critical question: how does the AI know when it doesn't know enough to act?

The answer lies in confidence scoring and user-set thresholds – mechanisms that determine when an AI agent should proceed autonomously and when it should stop and ask a human for help. For insurance professionals tasked with making AI purchase or deployment decisions, understanding how these systems work and how they can be configured is no longer optional knowledge.

May 2026: Insurance AI Trends & Highlights

Roots Experts — Thu, 21 May 2026 12:00:00 GMT

Here's your monthly digest of must-read insurance and AI updates, real-world use cases, and emerging innovations, with new articles added each week.

Latest articles as of May 21

News: AI claims need their own coverage category

The root of it: A viewpoint from former Everest cyber head Peter Hawley in Insurance Journal argues that insurers underwriting AI risk as a cyber problem are making a category error. The actual claims, Hawley writes, are not coming from breaches or ransomware but from routine customer calls, chatbot interactions, healthcare consultations, and vendor contract defaults that capture and process data in ways customers did not anticipate. Hawley points to the Invoca court's "mere capability" reading of California's Invasion of Privacy Act, which lowers the bar for plaintiffs to plead a claim. He argues the companies that understand the difference first will be ahead when the market corrects.

News: Cyber claim severity doubles for large accounts

The root of it: Chubb's 2026 Cyber Claims Report found average cyber claim severity for US businesses with $1 billion or more in revenue reached about $4.4 million in 2025, up from roughly $2.2 million in 2024. That is a 586% increase from 2021. Middle market severity rose to about $759,000, while SME severity actually fell. Chubb attributes the increase to business interruption costs and the rising cost of data breach and privacy litigation. Claim frequency declined across large accounts and middle market, even as severity climbed.

News: Specialty rates erase half the hard market

The root of it: WTW's latest Specialty Insurance Marketplace Survey found specialty rates in the London, Bermuda, and US E&S markets declined faster than expected at the January 1, 2026 renewals, with a 10-point drop on the rate index back to 2020 levels. About 45% of cumulative rate increase from 2017 to the 2023 peak has now been eroded over the past two years. 2025 was the first year since 2018 in which rate adequacy for new business was lower than for renewal business. Property and energy lines are showing the most pronounced decreases.

News: CEOs frame AI as rebalancing, not replacing

The root of it: Sollers Consulting's CEO Voices Report 2026 draws on interviews with 11 insurance CEOs including QBE's Andrew Horton, Markel Insurance's Simon Wilson, and Intact UK's Ken Norgrove. The report finds AI is moving from experimentation to operational use across the value chain, but executives reject the framing that AI inevitably eliminates jobs. Roles are shifting toward case management, advisory work, and portfolio steering, with reskilling and data literacy identified as the biggest near-term staffing priority. The CEOs interviewed lead carriers across seven countries.

News: Colorado replaces its landmark AI Act

The root of it: Governor Jared Polis signed SB 26-189, repealing and replacing Colorado's landmark 2024 AI Act weeks before its June 30 effective date. The new law drops the original risk management programs, impact assessments, and broad algorithmic discrimination duties in favor of a narrower notice and transparency framework focused on "automated decision-making technology" used in consequential decisions, with insurance explicitly named as a covered domain. SB 26-189 takes effect January 1, 2027. Baker McKenzie tells companies deploying ADMT tools to closely monitor forthcoming AG rulemaking and the status of stayed litigation challenging the prior law.

Latest articles as of May 14

News: Aon details how Middle East conflict is shifting underwriter behavior

The root of it: Aon says that the Middle East conflict has shifted volatility from a tail risk to a baseline planning assumption for insurers. Cynthia Beveridge, Aon’s Chief Broking Officer for Commercial Risk, said underwriter behavior is changing ahead of any direct loss activity, with sharper questions on exposure concentration, governance, and preparedness. Tony Day, Aon’s Crisis Management Leader, said insurers are re-engineering line size management within the Middle East market, making expiring limits harder to replicate. Cargo, transit values, and dwell times at ports are also getting closer scrutiny.

News: Q1 insurtech investment goes almost entirely to AI

The root of it: Gallagher Re’s Q1 2026 Global InsurTech Report found that AI-labeled insurtechs captured 95.2% of all venture funding in the first quarter, the highest concentration on record. The sector raised $1.63 billion total, with AI-focused firms taking $1.55 billion across 68 deals and all ten of the quarter’s largest rounds going to AI-led businesses. Companies in AI liability and cyber insurance raised $444.84 million. Gallagher Re’s Freddie Scarratt said third-party AI liability cover is on track to mirror the cyber reinsurance market’s growth trajectory.

News: Generalist brokers leave cyber clients structurally exposed

The root of it: Jennifer Wilson, head of cyber at Newfront, told Insurance Business Magazine that the cyber market remains constrained by inconsistent policy language and slow adaptation to AI-related exposures. Wilson said each carrier approaches cyber risk with its own definitions, limits, and exclusions, making policy comparison difficult. Policy revisions that once happened every few years are now happening quarterly. Wilson reads the speed of change as creating structural risk for clients placing cyber coverage through generalist brokers who may not track AI policy language closely.

News: Allianz transfers commercial cyber business to Coalition

The root of it: Allianz Commercial will transition its entire standalone commercial cyber book to Coalition under a 10-year exclusive partnership. Coalition will take over pricing, product development, risk mitigation, and claims management, while Allianz continues to provide capacity, distribution, and support for multinational accounts. The deal expands an existing relationship from 2022, when Allianz X first invested in Coalition. Allianz cites ransomware and AI-driven cyber-attacks as drivers behind consolidating on Coalition's Active Insurance platform.

News: PayPal lays out AI-driven operating model

The root of it: PayPal CEO Enrique Lores told investors on the Q1 earnings call that the company is “becoming a technology company again,” with plans to modernize the platform and “aggressively adopt AI in our development processes.” PayPal has formed a new “AI transformation and simplification” team to lead the work, which Lores said will drive at least $1.5 billion in cost savings over the next two to three years. Additional savings are expected to come from extending AI into customer service, support operations, and risk management

Latest articles as of May 7

News: White House weighing FDA-style vetting for AI models

The root of it: National Economic Council Director Kevin Hassett told Fox Business the White House is studying an executive order that would create a vetting system for new AI models, comparing the process to FDA drug approval. The push follows Anthropic's Mythos model, which can identify cyber vulnerabilities. For carriers writing cyber, E&O, and tech-related coverage, a federal pre-release safety regime could change how AI-related risk is rated and excluded. The story is early, with no executive order text yet released.

News: Capgemini names Roots AI a key example in its 19th annual P&C insurance report

The root of it: Capgemini's 19th annual World P&C Insurance Report — drawing on 344 senior executives, 809 employees, and 1,113 policyholders across the Americas, Europe, and Asia-Pacific — highlighted Roots and tapped Co-Founder & CEO Chaz Perera as an executive steering committee member alongside CIOs and CDOs from The Hartford, Intact, Markel, and Arch Insurance. The report identifies a small group of intelligence trailblazers achieving up to 21% higher revenue growth and 51% greater share price gains over three years by treating AI as a core operating capability.

News: AI seen as near-term underwriting risk for cyber insurers

The root of it: Fitch Ratings says US cyber insurers reversed two years of decline in 2025 to post 11% growth in direct written premiums, but flags AI as a near-term underwriting concern. Fitch points to Anthropic's Mythos model as a force that will likely create more vulnerabilities than patches in the short to medium term. Fitch wrote that AI is particularly disruptive to cyber risk because traditional vulnerability analysis was labor-intensive and offered limited financial upside for researchers, a gap AI now fills at scale. Volume drove most of the cyber market's growth, with policies-in-force up 35% offsetting soft pricing.

News: Carriers carve AI out of standard liability coverage

The root of it: Berkshire Hathaway, Chubb, and Travelers asked state regulators to exclude AI-related damages from general liability policies, and regulators approved more than 80% of those requests. Florida, Connecticut, and Maryland led approvals, with exclusions taking effect as early as January. Some carriers including Berkley introduced absolute AI exclusions across D&O, E&O, and cyber lines. Munich Re plus startups Corgi, Armilla, Mayflower Specialty, and Embroker are stepping in with standalone AI liability policies offering limits from $2 million to $50 million.

News: Insurer AI returns blocked by weak governance

The root of it: Grant Thornton's 2026 AI Impact Survey found 44% of insurance executives say governance or compliance challenges have contributed to AI projects failing or underperforming. Only 24% are very confident they could pass an independent AI governance review in 90 days. At the same time, 52% credit AI with revenue growth, 62% report better decision-making, and half say it has cut costs. About 61% of executives say their boards have established AI governance policies, but Grant Thornton notes the evidence and controls are often fragmented across teams and tools.

News: AI-manipulated photos showing up across claim types

The root of it: IA Magazine reports that AI-generated and manipulated content is showing up in claims across personal lines, commercial lines, and specialty coverages, from inflated home damage photos to staged car accidents and synthetic medical scans. Insureds are filing legitimate claims and using deepfake imagery to inflate damages, including AI-manipulated photos of luxury items. The article reinforces what Verisk has separately reported, that traditional review methods were not designed for synthetic media. IA Magazine writes that independent agents now have a more active role in fraud detection by verifying damage and contractor work directly with insureds.

What Insurers Need to Know about Colorado’s New AI Regulations

Roots Experts — Tue, 19 May 2026 16:23:13 GMT

Insurance companies have moved past asking whether to use AI. Now the focus is on making it work, getting AI embedded into core workflows and delivering real, measurable results.

The AI execution phase brings its own urgency, and with it, new exposures. The faster AI moves into underwriting, claims, and coverage decisions, the more it attracts regulatory attention. Carriers that treat compliance as something to sort out after deployment are learning that regulators aren't waiting.

Colorado made that point clear last fall. The state is the first in the country to specifically regulate how insurers use AI and external consumer data.

The Division of Insurance (DOI) put Regulation 10-1-1 into effect in October 2025, setting clear documentation and oversight requirements for life, private passenger auto, and health benefit plan insurers that use external consumer data, algorithms, and predictive models.

These regulations apply to any insurer authorized to do business in Colorado, regardless of where the company is headquartered. Now, any insurer writing life, auto, or health benefit plans in Colorado is subject to the same requirements as a Colorado-based insurer.

Build vs. Buy Insurance AI: Why You Can’t Prompt Your Way to Production

Roots Experts — Tue, 12 May 2026 15:44:05 GMT

Insurance leaders are under pressure to move faster with AI while controlling cost, risk, and complexity. Leadership expects technology delivered quickly, and not just prototypes, but solutions that work reliably in real production environments. Speed and accuracy are quickly becoming competitive advantages. Policyholders and agents expect immediate responses and consistent outcomes. When processes slow down or results are inconsistent, business is lost.

That pressure is driving a common question: should we build or buy insurance AI?

If my team can write prompts and use AI coding tools, why not build our own?

What Is a Large Language Model and How Does It Apply to Agentic AI?

Roots Experts — Tue, 05 May 2026 14:00:16 GMT

AI is changing how businesses operate, make decisions, and serve customers across insurance and virtually every industry. At the center of this shift is a technology called the large language model, or LLM. If you've used ChatGPT, Claude, Gemini, interacted with an AI-powered customer service tool, or noticed that Outlook is now suggesting entire sentences as you’re writing them, you've already encountered an LLM in action.

But what exactly is a large language model? And how does it connect to the more sophisticated concept of agentic AI that is increasingly driving enterprise transformation in insurance and elsewhere?

April 2026: Insurance AI Trends & Highlights

Roots Experts — Thu, 30 Apr 2026 15:00:00 GMT

Here's your curated list of important insurance and AI news updates, critical use cases, and the latest innovations to help you stay ahead of the curve.

Latest articles as of April 30

News: Insurers urged to prepare as AI adoption accelerates across the industry

The root of it: The AI insurance market is projected to grow from $8.6 billion in 2025 to nearly $60 billion by 2033. Eighty-six percent of insurers plan to spend more this year, and McKinsey found that early AI leaders generate six times the shareholder returns of slower adopters. Yet governance is lagging the spend. Only 24 percent of executives say their AI controls could survive an independent audit, even as 23 states plus DC have adopted the NAIC's model bulletin.

News: How the Hormuz crisis revealed insurance's data problem

The root of it: Hercules Systems CEO Alex Babin argues the marine market's scramble to reprice Hormuz war risk in days, not quarters, has exposed a structural weakness in delegated authority underwriting. Bordereaux still arrive as inconsistent PDFs and spreadsheets that AI can read but cannot verify against binding authority terms. Babin points to one vessel that transited a newly restricted zone and appeared in a coverholder's bordereaux three weeks later, priced at pre-escalation rates. His take: standardize at the source rather than layer more AI extraction on top of uncertain inputs.

News: New study puts 34 million coastal Americans at serious flood risk

The root of it: A new University of Alabama study in Science Advances finds 17.5 million people along the US Atlantic and Gulf coasts at "very high" flood risk and another 17 million at "high" risk. Researchers mapped the exposure using 16 factors and three AI tools across Texas to Maine.
The concentration is stark in some cities. New York alone has 4.75 million people at the two highest risk levels, and 99 percent of New Orleans' population falls into those top tiers. The authors point to natural infrastructure like wetlands as a complement to traditional dams and levees in the highest-risk corridors.

News: Beazley report finds businesses are overconfident in their cyber resilience as threats grow

The root of it: Beazley’s 2026 Risk and Resilience report, based on a survey of 3,500 business leaders worldwide, finds 78 percent believe they could fully recover financially from a cyber attack and 82 percent feel prepared, even as cyber holds the top spot in their concerns. Beazley reads the gap as overconfidence in a threat environment that is becoming more systemic. Eighty percent of respondents expect AI to improve financial performance while 72 percent now think it will replace jobs within 18 months, up from 66 percent a year ago.

News: AM Best finds insurers' AI ambitions are outrunning their readiness to execute

The root of it: A new AM Best survey of more than 150 rated insurers and managing general agents finds that 60 percent expect AI to significantly transform their business models within the next one to three years, while 41 percent are already using AI across core business areas. A majority report having a formal AI policy in place, with nearly 20 percent saying their organization has reached an advanced stage of implementation. Respondents identified data readiness, security, and integration with legacy systems as the largest impediments to wider deployment.

News: Re/insurers enter a costly year on solid footing after a mild Q1

The root of it: Aon and Gallagher Re both estimate Q1 2026 global insured natural catastrophe losses at roughly $20 billion, 26 percent below Gallagher Re’s decadal Q1 average and driven mostly by US winter storm and severe convective storm activity. Gallagher Re puts the threshold for meaningfully shifting property pricing at a single event or series totaling $115 to $125 billion in insured losses. The brokers also flagged a possible return to El Niño conditions by June or July, which historically suppresses Atlantic hurricane activity.

News: Ivans Index shows commercial lines rates continuing to soften

The root of it: The Q1 2026 Ivans Index shows premium renewal rates fell across every major commercial line compared to Q4 2025. Property dropped from 8.01 to 6.83 percent, commercial auto from 6.97 to 5.28 percent, and BOP from 7.52 to 6.74 percent. Workers' compensation extended its negative streak, coming in at minus 1.73 percent. Year-over-year rate change remains positive in every line except workers' comp. But the consistent quarter-over-quarter softening points to a deceleration trend that has now held through the first quarter.

Latest articles as of April 23

News: Gartner projects 2,000+ "death by AI" legal claims this year, signaling a new phase of AI liability risk insurers need to prepare for

The root of it: A Gartner Group report forecasts that more than 2,000 legal claims related to "death by AI" will be filed worldwide by year's end, as mistakes by AI software or those implementing it may be the root cause of fatalities. The implications are most immediate for health insurers, though the ripple effects will reach across the industry – particularly where AI is being used to anticipate and prevent losses. As juries have displayed the tendency to hold software to a higher standard than human judgment, insurers should brace for complex AI liability questions that the legal system has only begun to grapple with.

News:Insurers are vulnerable to many of the same cyber risks that they cover

The root of it: Recent research shows that, despite their role in establishing cybersecurity standards for policyholders, insurers still have significant vulnerabilities in their own defenses. The study flagged gaps in credential management, backup definitions, and patch deployment. While carriers generally follow robust security practices, some still allow less secure multi-factor authentication methods, and only about half deploy security patches monthly – a window that cyber criminals routinely exploit.

News: NYC announces the launch of a reduced-cost insurance program

The root of it: Mayor Zohran Mamdani's administration unveiled what it called a first-of-its-kind, city-backed insurance program aimed at lowering property and liability premiums for affordable housing and rent-stabilized buildings. The city plans to use its own financial resources to compete directly in the insurance market, driving down premiums by an estimated 20 to 30 percent. The program is expected to cover 20,000 homes by 2027 and 100,000 by 2030, addressing insurance costs that have tripled since 2017.

News: The Strait of Hormuz situation precipitates a "rethink" of marine insurance risk

The root of it: What began as a geopolitical flashpoint in the Strait of Hormuz is now a high-level insurance crisis, with shipowners and cargo clients demanding answers on coverage, costs, and continuity as vessels sit trapped in one of the world's most important maritime chokepoints. War risk pricing on some Hormuz transits climbed in March, from roughly 0.10% of vessel value before the conflict to around 2–3%. The crisis is also prompting a broader reassessment of other vulnerable sea trade corridors, including the Turkish Straits and the Malacca Strait.

News: Gallup poll says half of all US workers use AI on the job

The root of it: For the first time since Gallup began measuring workplace AI adoption statistics, half of employed American adults surveyed say they use AI in their role at least a few times a year – up from 46% in the last business quarter. While 65% of employees in organizations that have adopted AI report improved productivity, the gains remain concentrated at the individual task level. Only about one in ten polling participants “strongly agree” that AI has transformed how work gets done across their organization – suggesting that broad adoption has yet to drive fundamental workflow redesigns.

News: The parent company of Snapchat cuts 1,000 jobs due to "rapid advancements in AI"

The root of it: Snap is laying off roughly 16% of its global workforce, affecting around 1,000 full-time employees. The company cites advancements in AI as a reason for the cuts. AI now generates over 65% of new code at Snap, enabling the company to assign more critical work to focused teams and AI agents. The restructuring is expected to reduce Snap's annualized cost base by more than $500 million by the second half of 2026.

News: Three convicted in California “fake bear attack” auto insurance scam

The root of it: Three Southern California residents have been convicted of orchestrating a bizarre insurance fraud scheme that involved staging fake bear attacks on luxury cars for insurance payouts. Dubbed "Operation Bear Claw," the scheme involved filing claims alleging a bear had destroyed high-end vehicles in the Lake Arrowhead area – with video evidence of a person in a bear costume – ultimately collecting over $141,000 in fraudulent payouts. The three defendants received jail time, probation, and restitution orders; a fourth is due in court in September.

Latest articles as of April 16

News: Willis launches data center coverage, risk management programs

The root of it: Willis, a WTW business, launched Digital Infrastructure Protector on April 9, an end-to-end lifecycle solution for data center owners, operators, contractors, and hyperscalers. Developed in collaboration with Zurich, the product provides more than $3 billion in integrated capacity, bringing building, operational property, marine, and cargo exposures under a single policy. An eight-point risk framework delivers ongoing guidance as projects evolve, while evidence-based broking helps clients avoid overinsurance.

News: Reinsurance rates continue to soften despite Middle East instability

The root of it: Reinsurance rates continued easing at the April 1 renewal, with Aon, Gallagher Re, Guy Carpenter, and Howden all reporting generally buyer-friendly outcomes. Robust reinsurer balance sheets, abundant capacity, and benign catastrophe losses in 2025 and Q1 2026 drove the trend. The Iran conflict's impact was confined to marine war risk, energy, and political violence specialty lines — property-catastrophe renewals proceeded without structural disruption.

News: Is the North Atlantictrending away from “moderate” hurricanes?

The root of it: Colorado State University's annual Atlantic Hurricane Forecast, released April 9, calls for a slightly below-average 2026 Atlantic hurricane season, with 13 named storms, six hurricanes, and just two major hurricanes, activity driven by the anticipated transition to moderate-to-strong El Niño conditions. Landfall probabilities are expected to be at their lowest levels since 2015. CSU and other forecasters caution that a single landfalling storm can make any season consequential regardless of overall activity.

News: Harvard economists: AI job risk is a much bigger economic disruptor than tariff shock

The root of it: A panel of Harvard economists, including former IMF deputy chief Gita Gopinath, warned that AI's threat to labor markets is "of a much bigger magnitude" than tariff disruption, with roughly 30% of jobs in advanced economies potentially vulnerable. Gopinath raised the risk of a recession coinciding with AI-driven displacement – a combination that could produce job losses exceeding those of the 2008 financial crisis. Panelists also flagged the possibility of an AI investment bubble.

News: Hormuz crisis exposes limits of private war-risk insurance at scale

The root of it: The Iran conflict has effectively halted shipping through the Strait of Hormuz, with Lloyd's expanding its high-risk designation to the entire Persian Gulf and major Protection and Indemnity (P&I) entities canceling war-risk cover. Tanker traffic fell by roughly 95%. The US government stepped in with a $40 billion Development Finance Company (DFC) reinsurance facility – raising broader questions about the boundaries of private insurability when geopolitical risk escalates rapidly, and governments become insurers of last resort.

News: Meta banssocial media addiction litigation advertisingfrom its platforms

Following back-to-back jury defeats in California and New Mexico, Meta began removing law firm advertisements from Facebook and Instagram that sought to recruit plaintiffs for social media addiction lawsuits. Axios identified more than a dozen deactivated ads from major national firms. Meta stated it would not allow trial lawyers to profit from its platforms while simultaneously claiming they are harmful. Thousands of similar lawsuits remain pending in state and federal courts.

News: Alibaba surprises as the new leader in dedicated AI video generation capabilities

The root of it: Alibaba's HappyHorse1.0, launched anonymously on April 7 and confirmed on April 10, reached the top of the Artificial Analysis global leaderboard for both text-to-video and image-to-video generation – the most significant benchmark margin reached in the platform's history. The 15-billion-parameter model, built by Alibaba's new ATH AI Innovation Unit, generates 1080p video with synchronized audio in a single pass across seven languages. Its rise coincides with OpenAI's exit from video generation and ByteDance's pause over copyright-related issues.

Latest Articles as of April 9

News: Traditional insurance unprepared for AI liability, per Gallagher Re/MIT report

The root of it: Generative AI-related lawsuits in the US grew 978% between 2020 and 2025, yet standard insurance policies – cyber, tech E&O, product liability, and commercial general liability – each leave significant coverage gaps, according to a Gallagher Re report developed with MIT. Courts are treating AI as a tool, placing liability on the organizations deploying it. Vendor contracts compound the problem, typically capping liability at 12 months of fees with no performance warranties.

News: US E&S market growth slows to lowest rate in 8 years

The root of it: The US excess and surplus lines market crossed $100 billion in direct premiums written for the first time in 2025, reaching $105.31 B – but growth decelerated to 7.8%, the lowest rate since 2018, according to S&P Global Market Intelligence. Declining commercial property premiums drove the slowdown, with combined property lines falling 2.8%. E&S homeowners coverage surged 29.5%, its third consecutive year of 20%-plus growth, as admitted carriers continued retreating from catastrophe-prone states.

News: AI explainability emerges as the new standard for reducing regulatory risk

The root of it: As insurers deploy AI across underwriting and claims, the bar for accountability is shifting from "we trust the model" to "we can defend the outcome," according to insurance professionals interviewed by InsuranceNewsNet. Explainability now means being able to walk a regulator, broker, or policyholder through any pricing or coverage decision in plain English. Carriers that embed explainability and auditability from day one are positioned to outperform those treating AI as a black-box shortcut.

News: AI’s role in organizational decision-making is redefining exposure

The root of it: Rather than representing a single, discrete exposure, AI amplifies existing risks across cyber, professional services, employment, intellectual property, product liability, and D&O, according to Aon's AI Risk 2026: What Business Leaders Need to Know report. AI-generated phishing now achieves click-through rates of roughly 54%, compared with 12% for traditional attacks. More than 90% of insurance decision-makers consider AI-driven incidents a material concern. Boards face rising D&O scrutiny over AI governance, and underwriters increasingly factor governance maturity into policy terms and capacity.

News: Judge orders UnitedHealth to provide AI claim denial documents by April 29

The root of it: A federal court in Minnesota has ordered UnitedHealth Group to produce tens of thousands of internal documents by April 29 in a lawsuit alleging the company used an AI algorithm, known to have a 90% appeal reversal rate, to systematically deny Medicare Advantage claims. The discovery order covers nearly a decade of AI governance records, cost-saving analyses, and performance metrics. Legal analysts say the documents, not the trial itself, may reshape how healthcare organizations use AI in coverage decisions.

News: Private capital-fueled data center boom creates new challenges and opportunities for insurers

The root of it: Insuring a $20 billion AI data center campus was nearly impossible in 2023; today, it's a regular conversation for specialist brokers. Global AI data center spending could reach $7 trillion by 2030, with Big Tech increasingly turning to private equity and debt markets to fund construction. The concentration of value, bleeding-edge technology, and opaque financing structures presents both capacity challenges and significant new opportunities for insurers willing to build dedicated expertise to serve the market.

News: Neuro-symbolic AI offers promise of greater intelligence and significantly lower energy consumption

The root of it: Researchers at Tufts University have developed a neuro-symbolic AI system that combines neural networks with human-like symbolic reasoning, achieving a 95% success rate on structured tasks compared with 34% for standard systems. Training required only 1% of the energy presently used by conventional models – and operation consumed just 5%. AI already accounts for more than 10% of US electricity demand, a figure projected to double by 2030, making more efficient architectures an urgent area of research.

Latest Articles as of April 2

News: China's largest private insurer expects to reap a $174B benefit from AI claims automation

The root of it: Five years ago, nearly all accident and health claims at Ping An Insurance Group required human intervention. Today, nearly 60% are automated – with some settled in as little as 51 seconds. A decade of AI investment is also reshaping the firm's financial ambitions: Executives are banking on the technology to double Ping An's price-to-book ratio, which would add approximately $174 billion to the market value of China's largest non-state-owned insurer.

News: WTW survey finds lower combined ratios and higher premium growth for carriers using AI and advanced analytics

The root of it: P&C insurers that invested more heavily in advanced analytics and AI outperformed slower adopters between 2022 and 2024, achieving combined ratios six points lower and premium growth three points higher, according to WTW's 2026 Advanced Analytics and AI Survey. Claims functions lag in adoption, but that's changing fast – the share of insurers using analytics for fraud detection is expected to roughly double over the next two years.

News: AI poised to turn "niche" risks into new specialty insurance markets

The root of it: Andrew Kelly, executive vice president at AJ Wayne & Associates, an E&S lines wholesaler/MGA, sees AI liability evolving the way cyber did in the 1990s – from a niche exposure into a major standalone line. Kelly predicts a dedicated AI insurance sector could emerge within five to ten years, complete with its own MGAs, claims professionals, and policy forms. For now, most policies address AI exposures only indirectly, through endorsements or gaps in existing language.

News: Traditional insurance is unprepared for AI liability exposures, according to a Gallagher Re study

The root of it: Generative AI-related lawsuits in the U.S. grew 978% between 2021 and 2025, yet standard insurance policies – cyber, tech E&O, product liability, and commercial general liability – each leave significant gaps in coverage, according to a Gallagher Re report produced with the Massachusetts Institute of Technology. Courts generally treat AI as a tool, placing liability on organizations that deploy it. Vendor contracts compound the problem, typically capping liability at 12 months of fees with no performance warranties.

News:Hail risk could deliver a billion-dollar blow to The Windy City

The root of it: Chicago isn't typically considered a major “hail market” – but more than 1.7 million homes there face moderate or severe hail risk, representing over $1 trillion in reconstruction value, according to a new Cotality report. Illinois ranks second nationally in hail exposure, behind only Texas. Cotality's most extreme modeled scenario puts potential losses from a single severe hailstorm at $58 billion – roughly equivalent to the insured losses from 2022’s Hurricane Ian.

News:Court temporarily blocks Pentagon's "national security risk" declaration against Anthropic

The root of it: A federal judge in California granted Anthropic a preliminary injunction blocking the Trump administration's ban on its technology and the Pentagon's designation of the company as a supply-chain risk. U.S. District Judge Rita Lin found the government's actions appeared designed to punish Anthropic for publicly criticizing the DOD's contracting position – ruling that the company had adequately demonstrated the measures were likely unlawful and that it was suffering irreparable harm.

News: After massive social media addiction liability verdicts, what comes next for Meta, et al?

The root of it: A Los Angeles jury ruled that Instagram and YouTube are deliberately engineered to be addictive and that Meta and Google were negligent in protecting children who used them – ordering $6 million in damages to a plaintiff who claimed the platforms caused body dysmorphia, depression, and suicidal thoughts. It was big tech's second such defeat this year, with more trials ahead. Some experts are calling it the industry's "big tobacco" moment.

How to Know If an Insurance AI Solution Is Actually Good

Roots Experts — Tue, 28 Apr 2026 16:14:26 GMT

You are researching AI solutions for your insurance operation. You have read the articles, maybe attended a webinar or two, and now you are sitting through vendor demos. Everything looks impressive. The interfaces are clean. The use cases sound familiar. The sales team is confident.

And then a quiet question forms in the back of your mind. "But how do I actually know if this AI solution is good?"

Vendors sell outcomes. What you need to evaluate is the AI itself. Is it truly capable of delivering on vendor promises? Will it hold up under real conditions?

Here are key areas that tell you whether the AI behind the solution you’re evaluating is genuinely good or just well packaged.

93% of Insurers Are Now Actively Using or Exploring AI

Roots Experts — Tue, 21 Apr 2026 14:23:59 GMT

It's Monday morning. Your underwriting team is staring down 200 submissions that came in over the weekend. Your claims unit just got hit with a new CAT event. And somewhere in a conference room, a steering committee is still debating which AI vendor to shortlist. Meanwhile, 93% of the industry has already moved past that conversation. The gap isn't coming. It's here.

Based on Roots' newly released State of AI Adoption in Insurance 2026 report, the question is no longer whether AI is relevant. It's where to deploy it next, how fast to scale, and how to build the governance structures that make responsible deployment possible. This is a market that has largely made up its mind about AI's value. The debate has shifted from "should we?" to "how do we do this well?"

How Insurers Are Using AI to Summarize Claims Faster and More Accurately

Roots Experts — Tue, 14 Apr 2026 16:09:45 GMT

A claims summary isn't a single document with a single purpose. It's what an adjuster produces when updating a policyholder on where their claim stands, when bringing a broker or agent into the conversation, and when transferring a file to another handler. Each of those moments calls for a summary, and each one looks different depending on who's receiving it.

A policyholder needs to know what happened and what comes next. A broker needs enough context to respond to their client. An incoming adjuster needs a full picture of activity, documentation, and outstanding items. The same claim, three different summaries.

That variation is what makes this work time-consuming. Adjusters aren't just writing a recap. They're pulling together police reports, medical records, activity logs, and adjuster notes, then deciding what matters most for this audience, at this stage of the claim. Multiply that across a caseload of 150 to 200 open files, and the time adds up fast.

AI-powered claims summarization addresses this directly. Instead of requiring adjusters to manually piece together each summary, an AI agent analyzes the full claim file and generates a clear, structured output in minutes, tailored to the audience and the moment.

The Build Trap: Why AI Is Repeating Insurance’s Oldest Technology Mistake

Roots Experts — Tue, 07 Apr 2026 18:29:52 GMT

Insurance has always built its own technology. When there were no vendors, that was a necessity. When vendors arrived, it became a habit. Now AI is here, and that habit is harder than ever to break.

While building solutions in-house has never seemed more attainable, sustaining and improving what you build has become far more difficult and complex. Before deciding whether to build or buy AI systems, let’s take a moment to understand why this gap exists and how to avoid it.