AI Liability Insurance Gap 2026: Why Major Insurers Are Dropping AI Coverage and What Your Business Must Do Now
Photo by Winston Chen on Unsplash
- Major insurers — including AIG, Berkshire Hathaway, Chubb, and Travelers — have filed to strip AI-related liabilities from standard commercial policies, with state regulators approving more than 80% of those requests.
- New Verisk/ISO endorsement forms (CG 40 47 and CG 40 48), effective January 1, 2026, give carriers a standardized tool to exclude generative AI claims from general liability policies — policies renewing in Q1 and Q2 2026 are the first materially affected.
- A standalone AI liability insurance market is emerging, with Munich Re, Armilla, Corgi, Mayflower Specialty, and Embroker offering policy coverage limits ranging from $2 million to $50 million per organization.
- Any business using AI tools — chatbots, automated agents, AI-generated content — should audit its current policy coverage before its next renewal date.
What Happened
In November 2025, three of the largest names in U.S. commercial insurance — AIG, Great American, and W.R. Berkley — filed requests with state regulators to do something that would have seemed unlikely just a few years ago: formally remove coverage for AI-related liabilities from their standard policies. Specifically, they asked to exclude harms caused by AI tools such as chatbots and autonomous agents. By early 2026, major carriers including Berkshire Hathaway, Chubb, and Travelers had joined the movement.
State regulators have approved more than 80% of these exclusion requests, making this one of the fastest-moving coverage shifts in recent insurance history. Policies renewing in Q1 and Q2 2026 are the first to be materially affected — meaning many business owners are only discovering the gap when renewal documents land in their inbox.
To help carriers implement these exclusions consistently, Verisk/ISO — a leading insurance data and analytics organization — introduced two new standardized endorsement forms (add-ons that modify a base policy) on January 1, 2026: CG 40 47 and CG 40 48. These forms define generative AI as any "machine-based learning system or model trained on data with the ability to create content or responses, including text, images, audio, video, or code." In plain English: if your AI chatbot gives a customer harmful advice, your automated agent makes a costly operational error, or your AI-generated content triggers a lawsuit, your standard commercial general liability (CGL) policy likely won't cover the fallout.
The exclusions don't stop at CGL. They're spreading across Directors and Officers (D&O) coverage — which protects executives from personal liability for company decisions — Errors and Omissions (E&O) coverage — which covers professional mistakes — and cyber insurance lines as well. The ripple effects are broad and, for many business owners, unexpected.
Photo by Belle Lee on Unsplash
Why It Matters for Your Coverage
Here's an analogy that might hit close to home: imagine you've operated a small accounting firm for years, and your professional liability policy has always covered mistakes made by your staff. Now imagine your insurer quietly added an exclusion for any mistake made using an AI-powered tax software tool — without flagging it until renewal time. That is effectively the situation thousands of businesses using AI tools are facing in 2026.
This isn't a niche problem affecting only tech companies. If your business uses an AI chatbot to handle customer inquiries, an AI tool to write marketing content, or an automated system to manage scheduling or billing, your policy coverage may now have a material gap. Karthik Ramakrishnan, founder and CEO of Armilla — one of the leading standalone AI liability carriers — put it plainly in December 2025: "This is a coming of age for the AI market — exclusions were inevitable because commercial general liability policies were never designed to cover AI."
For a meaningful insurance comparison, you need to understand what's actually being excluded. The Verisk/ISO CG 40 47 and CG 40 48 endorsement forms are intentionally broad. They don't target only enterprise AI platforms — they cover virtually any machine-learning-powered tool that generates outputs. That includes everyday business tools: AI email assistants, customer service bots, automated invoice processors, and even AI-assisted claims management platforms your insurance provider may use on its own end.
The financial stakes driving this shift are significant. Industry analysts warn that insurers aren't primarily afraid of one large, isolated claim — they're afraid of systemic, correlated loss (losses that hit many policyholders simultaneously from a single source). As one reinsurance market observer explained: "The industry could absorb a $400 million or $500 million hit from a misfiring agent used by one company, but it cannot absorb an upstream failure that produces a thousand simultaneous losses from a single model or vendor failure." This mirrors catastrophe exposure — a hurricane doesn't damage just one building; it damages thousands at once. AI risk follows the same pattern, and the industry has no historical actuarial data to price it reliably.
This is precisely why a standalone AI liability market is emerging. Insurance Intel has sized the opportunity at approximately $4.7 billion. Carriers like Armilla have already expanded their AI Liability Policy to cover up to $25 million per organization, including hallucinations (when an AI generates confidently wrong information), model drift (when AI performance degrades over time without detection), inaccurate outputs, data leakage, and AI regulatory violations. Other carriers — Munich Re, Corgi, Mayflower Specialty, and Embroker — are offering limits ranging from $2 million to $50 million per policy.
Not everything is gloomy on the broader insurance front. Fitch Ratings reported that U.S. cyber insurers reversed two years of premium decline in 2025, posting 11% growth in direct written premiums — even while flagging AI as a near-term underwriting concern. The insurance industry sees AI as both a commercial opportunity and a serious risk management challenge at the same time. For businesses thinking about insurance savings, understanding this landscape now — before your next renewal — is far cheaper than discovering a gap after a claim.
The AI Angle
The same AI tools creating policy coverage headaches for policyholders are simultaneously transforming how insurers handle claims management and underwriting internally. Most major carriers now deploy AI-driven systems to triage incoming claims, detect fraud, and accelerate routine settlements. Tools like Tractable (used for AI-powered auto damage assessment) and Shift Technology (used for fraud detection and claims management automation) have become standard parts of the insurance workflow.
The irony is hard to miss: carriers are excluding AI-related liabilities from policies while doubling down on AI for their own internal operations. This dual dynamic is accelerating demand for clearer AI accountability standards across the entire insurance value chain.
From an underwriting perspective, the retreat from AI coverage is pushing carriers to build better risk assessment models specifically for AI-driven exposures. Insurtech platforms like Armilla and Corgi are developing proprietary scoring tools to evaluate a business's AI maturity, governance practices, and liability exposure — essentially creating a new discipline of AI-specific underwriting that didn't exist 18 months ago. Your AI governance practices may soon matter as much to an underwriter as your credit score matters to a lender.
What Should You Do? 3 Action Steps
Request your complete policy documents and all endorsements from your insurer or broker before your next renewal. Look specifically for language referencing "generative AI," "machine learning," "automated systems," or the Verisk/ISO form numbers CG 40 47 and CG 40 48. Don't assume your policy coverage is unchanged just because your premium stayed the same — exclusion endorsements can be added at renewal without a rate change. Ask your broker directly: "Has any AI exclusion language been added to my policy?"
If your business uses AI tools in any customer-facing or operational capacity, conduct a thorough insurance comparison across the standalone AI liability market. Munich Re, Armilla, Embroker, Corgi, and Mayflower Specialty are currently active in this space, with limits from $2 million to $50 million per policy. Pricing is based on factors including which AI tools you use, how you govern and audit them, what data they handle, and your industry. Don't wait for a claim to prompt this conversation — the standalone AI market is new, and underwriting capacity is limited.
The intersection of AI liability, cyber, D&O, E&O, and CGL coverage is genuinely complex. A broker specializing in commercial lines and technology risks can map your specific AI tool usage to your current policy coverage, identify gaps in your risk assessment profile, and design a layered coverage strategy that addresses your actual exposure. This is not a do-it-yourself exercise. Exclusion language is technical, the standalone AI market is evolving rapidly, and the insurance savings from closing a coverage gap before a claim will almost always far outweigh the cost of a specialist consultation. Always work with a licensed agent for guidance tailored to your situation.
Frequently Asked Questions
Does my business general liability insurance still cover AI chatbot errors after the 2026 exclusions take effect?
As of early 2026, likely not — at least not if your carrier has adopted the new Verisk/ISO CG 40 47 or CG 40 48 endorsement forms. Major carriers including Berkshire Hathaway, Chubb, Travelers, and W.R. Berkley have filed to exclude AI-related damages from general liability (CGL) policies, and state regulators have approved more than 80% of those requests. Policies renewing in Q1 and Q2 2026 are the first to be materially affected. If your business uses any AI chatbot for customer service or communications, review your renewal documents carefully or ask your broker about specific AI exclusion language on your current policy.
What does standalone AI liability insurance cover and how much does it cost in 2026?
Standalone AI liability insurance is a new type of specialty policy designed specifically to fill the gap left by mainstream exclusions. Coverage typically includes harms caused by AI hallucinations (confidently wrong outputs), model drift (performance degradation over time), inaccurate AI outputs, data leakage, and AI regulatory violations. Carriers like Munich Re, Armilla, Corgi, Mayflower Specialty, and Embroker currently offer limits from $2 million to $50 million per policy. Armilla's policy, for example, covers up to $25 million per organization. Pricing varies based on your industry, AI tools used, governance practices, and data handling. A licensed commercial insurance broker can provide a specific quote based on your risk profile.
How do the Verisk ISO CG 40 47 and CG 40 48 endorsements change my existing commercial insurance policy?
These are standardized endorsement forms — modifications added to a base commercial general liability policy — introduced by Verisk/ISO on January 1, 2026. When attached to your policy at renewal, they exclude coverage for any harm caused by generative AI, defined broadly as any machine-based learning system capable of producing text, images, audio, video, or code. In practical terms, this means AI-related claims — even incidental ones involving tools you didn't consider "AI" — may be denied under your policy coverage. Ask your broker at your next renewal whether either form has been attached to your policy.
Can my small business face a claims management disaster if we use AI tools without standalone coverage in 2026?
Yes, and this is precisely the scenario insurance professionals are warning about. If a customer suffers harm from an AI-generated error — a chatbot giving incorrect product information, an automated system mishandling personal data, or an AI-drafted document containing a costly mistake — your standard CGL, E&O, or cyber policy may explicitly exclude the claim under new 2026 endorsement language. The claims management implications are serious: legal defense costs alone can run into six figures even for claims that are ultimately dismissed. A risk assessment of your AI tool usage with a licensed broker is the critical first step to understanding your true exposure.
Are there insurance savings available for businesses that show strong AI governance practices when buying AI liability coverage in 2026?
Potentially, yes. Standalone AI liability carriers are building underwriting models that reward businesses demonstrating strong AI governance — including regular model testing, documented audit trails, human oversight protocols, and AI usage policies. Much like how strong cybersecurity hygiene can lower cyber insurance premiums, your AI governance practices may affect both your eligibility for standalone coverage and your premium costs. Conducting a proper insurance comparison across multiple AI liability carriers and presenting documented governance evidence could yield meaningful insurance savings over time. Consult a licensed agent to understand how underwriters are currently evaluating these factors.
Explore Our Network
Disclaimer: This article is for informational purposes only and does not constitute insurance advice. Coverage terms, exclusions, and availability vary by carrier, state, and policy. Always consult a licensed insurance agent or broker for personalized guidance tailored to your specific situation.
