Most businesses that use generative AI tools assume their existing business insurance would respond if something went wrong. As of January 2026, that assumption is increasingly incorrect.
New industry-wide policy endorsements now allow insurers to formally exclude generative AI claims from standard commercial general liability policies. Similar exclusions are appearing on cyber liability, errors and omissions, and professional liability policies. The exclusions are being adopted quickly by major carriers, and they are showing up quietly on renewal policies without the fanfare that a change of this magnitude would normally receive.
The result is a growing gap between what businesses assume is covered and what their policies will actually respond to when a generative AI-related claim is filed.
This article walks through what the new exclusions do, why they exist, which businesses are most exposed, and how to actually close the coverage gap before renewal. For a broader overview of business coverage in general, our article on what is liability insurance for small businesses provides useful context.
What Changed in January 2026
The Insurance Services Office, commonly called ISO, is the industry organization that drafts standard policy forms used by the majority of U.S. property and casualty carriers. When ISO releases new endorsement language, it typically becomes widely adopted across the industry within a year or two.
In January 2026, ISO released three new endorsements specifically designed to exclude generative AI claims from commercial coverage. According to reporting from Insurance Journal and the Independent Agent Magazine, the three endorsements are:
CG 40 47
The broadest of the three. Excludes coverage under Coverage A (bodily injury and property damage) and Coverage B (personal and advertising injury) for any claims arising out of generative artificial intelligence. This is a full carve-out from standard general liability protection.
CG 40 48
Narrower. Excludes only Coverage B (personal and advertising injury) claims arising out of generative AI. This preserves potential coverage for bodily injury and property damage claims but eliminates coverage for defamation, copyright, trademark, and similar claims tied to AI outputs.
CG 35 08
Applies to products and completed operations coverage. Excludes bodily injury and property damage arising out of generative AI within that coverage section.
These endorsements are optional at the carrier level, but adoption has been rapid. According to industry analysis published by actuary.info, major carriers including W.R. Berkley, Chubb, Travelers, Berkshire Hathaway, and Cincinnati Financial had filed to adopt these endorsements or their own proprietary AI exclusion language by April 2026, with state regulators approving more than 80 percent of submitted filings.
Why Insurers Introduced These Exclusions
The move follows a pattern the insurance industry has seen before. In the late 2010s, cyber-related claims began appearing under general liability policies that had never anticipated cyber exposure. Insurers refer to this as “silent coverage,” where a policy responds to a category of loss the underwriters never priced for.
The industry’s response then was to add specific cyber exclusions to general liability and expand standalone cyber insurance. The same pattern is now playing out with generative AI.
According to legal analysis from PolicyholderPulse, AI-related lawsuits have grown rapidly. Copyright infringement claims tied to training data, defamation claims from AI-generated content, discrimination claims from AI-driven decisions, and property or bodily injury claims from AI-influenced errors are all showing up in litigation. Insurers do not want their traditional coverage lines absorbing an emerging risk that was never underwritten or priced.
The endorsements are the industry’s response: convert silent AI coverage into an explicit exclusion, and require businesses that need AI protection to purchase separate affirmative coverage.
What Kinds of Claims Are Now Excluded
The scope of the exclusions is broad. Any claim that has a meaningful connection to generative AI can potentially fall within the exclusion, regardless of whether the AI was used directly by the business or indirectly through vendors, consultants, or third-party tools.
Specific examples of claims that may now be excluded include:
- AI-generated marketing content that defames a competitor or third party
- AI-generated images, video, or written content that infringes copyright or trademark
- AI-driven hiring, screening, or evaluation tools alleged to discriminate against protected classes
- AI-generated customer service responses that misrepresent products or terms
- AI recommendations that lead a customer to a financial or personal loss
- AI-assisted design or engineering work that contributes to property damage or bodily injury
- AI-generated code that produces software failures or security vulnerabilities
- AI-driven autonomous systems that cause physical harm
The language in the endorsements is intentionally broad. Some proprietary carrier exclusions go even further, using phrasing like “based upon, arising out of, or in any way involving” generative AI. Under that kind of language, even loose or indirect connections to AI use can potentially trigger the exclusion.
Which Businesses Are Most Exposed
The exclusions apply to any business that uses generative AI, directly or indirectly. That is a significantly larger group than most owners realize.
Direct users include businesses that have integrated AI tools into their marketing, customer service, product design, code development, hiring processes, or client-facing communications. Any business using ChatGPT, Claude, Midjourney, Copilot, or similar tools falls into this category.
Indirect users are exposed through the tools and vendors they rely on. If your marketing agency uses AI to draft copy for you, if your customer service platform uses AI for responses, if your recruiting software uses AI to screen candidates, or if your software vendor uses AI in their product, the exclusions can still apply to claims that reach your business.
Some industries carry particularly acute exposure:
- Marketing agencies and content creators using AI to generate deliverables
- Technology companies embedding AI into products or services
- Professional services firms using AI for client work (accounting, consulting, legal support)
- Healthcare organizations using AI for scheduling, triage, or clinical support
- Financial services firms using AI for advice, screening, or communications
- Construction and engineering firms using AI for design, estimating, or safety analysis
- Retail and e-commerce businesses using AI for personalization, pricing, or recommendations
- HR technology providers and businesses using AI-based hiring tools
Legal alerts from Gallagher specifically flag construction organizations as at risk, since AI-enabled tools are increasingly embedded in estimating, scheduling, safety, and project coordination. The exclusion applies whether the AI is used directly by the insured or indirectly through vendors and project systems.
The Interaction With Cyber, D&O, and Professional Liability Policies
The general liability exclusions are only part of the picture. Similar exclusions are appearing on other commercial policy types, which limits the alternative places a claim might be picked up.
Cyber Liability Policies
Some cyber carriers are adding their own AI-specific exclusions or narrowing coverage for AI-driven incidents. Others are treating AI as a covered exposure but reviewing individual policy language carefully. Cyber policies were not designed to respond to AI hallucinations, output errors, or IP infringement claims, and the industry is still working out how to categorize these losses.
Directors and Officers (D&O) Policies
Shareholder claims tied to AI implementation, disclosures, or missteps can potentially trigger D&O coverage. However, some D&O carriers are also adding AI-related exclusions or narrowing certain provisions.
Errors and Omissions (E&O) or Professional Liability
For service businesses, professional liability is the most obvious backup when general liability excludes AI. But E&O policies from many carriers are also adding AI-specific exclusions or endorsements that narrow coverage. Even without a specific exclusion, some E&O policies may not respond to AI-driven errors because the policy language requires the error to be made by a person providing professional services.
The combined effect is that a business relying on general liability, cyber, and E&O together may find that all three now exclude the same category of claims. The gap between coverages becomes wider rather than narrower.
What Affirmative AI Coverage Looks Like
Because standard coverages are increasingly excluding AI risks, the insurance industry has begun developing affirmative AI coverage, which is coverage specifically designed to respond to AI-related claims.
Affirmative AI coverage generally takes two forms:
AI Write-Back Endorsements
These are optional endorsements that add specific AI coverage back to a general liability, cyber, or E&O policy. The scope varies by carrier, but they typically cover some or all of the claim types the base policy excludes. Write-back endorsements are the most common way affirmative coverage is being delivered right now.
Standalone AI Liability Policies
Some carriers have introduced standalone AI liability policies designed specifically to cover generative AI exposures. These are still emerging and vary significantly in scope and pricing. Coverage typically includes third-party bodily injury, property damage, personal and advertising injury, and errors and omissions arising from AI use.
The market for affirmative AI coverage is still developing. Availability, pricing, and terms vary significantly. Some coverage options exclude specific industries or restrict use cases. Others require the business to demonstrate specific AI governance practices as a condition of coverage.
What Businesses Should Do Right Now
The gap between what businesses assume they have and what their policies actually cover is likely to widen as 2026 renewals continue. Several concrete steps can help close the gap before it produces an uncovered claim.
Pull Your Current Policies and Read the Endorsements
Locate your current commercial general liability, cyber, D&O, and E&O policies. Look specifically at the endorsement schedule. Search for any language referencing generative AI, artificial intelligence, or machine learning. If you find one of the ISO endorsements (CG 40 47, CG 40 48, or CG 35 08) or proprietary equivalents, the exclusion is already in place.
Identify Every Place You Use AI
Take a systematic inventory of how AI is used in your business. Include direct use by employees, use by third-party vendors, and use embedded in software you rely on. Any of these can create exposure. The inventory is the starting point for a coverage conversation.
Ask Your Insurance Advisor Specific Questions
Do not accept general reassurances. The specific questions to ask include: does my current policy include any AI exclusions, what affirmative AI coverage options are available for my industry, how does the exclusion interact across my general liability, cyber, and E&O policies, and what is the additional cost to close the gap.
Review Vendor Contracts
If your vendors use AI in the services they provide to you, review whether their contracts include indemnification or insurance requirements that protect you if AI-driven work causes a claim against your business. Contracts written before AI exclusions became common may not provide the protection you assumed they did.
Document Your AI Governance
Businesses that can demonstrate structured AI governance, including approved use cases, human review requirements, and clear documentation of AI-generated outputs, are in a stronger position to obtain affirmative coverage and to defend claims when they arise. Governance is becoming an underwriting factor.
These steps parallel the review process for any commercial insurance program. Our article on business insurance mistakes small companies make covers the broader set of coverage gaps that can leave businesses exposed.
Why This Matters More Now
The pace of AI adoption in business has outrun the pace of most companies updating their insurance programs to match. Businesses are using AI tools in customer-facing, product-facing, and decision-making processes without necessarily thinking through the liability implications.
At the same time, plaintiff attorneys have begun building expertise in AI-related litigation. The types of claims that can arise from AI use are already being tested in court. As more precedent develops, the frequency and cost of AI-related claims is expected to rise.
The businesses that will fare best in this environment are the ones that address the coverage gap proactively rather than waiting to discover it in a claim. The ones that get caught flat-footed will be the ones whose renewal added the exclusion without them noticing, and who then face a claim they assumed their policy would handle.
For businesses concerned about legal exposure more broadly, our article on how to protect your small business from a lawsuit covers the layered approach that reduces overall litigation risk, including through proper insurance structure.
Getting Ahead of the Gap Before It Costs You
The introduction of AI exclusions in standard commercial policies is one of the most significant coverage shifts the insurance industry has seen in years. It affects businesses across nearly every sector, and the effects will only grow as more carriers adopt the endorsements and more claims start hitting the exclusion.
The right response is not to avoid AI. That train has left the station for most businesses. The right response is to understand exactly what your policies cover, what they exclude, and how to structure coverage that matches the way your business actually operates.
If you are ready to review your current commercial coverage or evaluate affirmative AI insurance options, InsuranceHub’s business insurance team can help you audit your existing policies, identify any AI exclusions in place, and explore the coverage options that are emerging in the market. You can also request a quote to start the conversation.
Frequently Asked Questions
What are the new AI exclusions in commercial insurance policies?
In January 2026, the Insurance Services Office released three new endorsements (CG 40 47, CG 40 48, and CG 35 08) that allow insurers to exclude claims arising out of generative artificial intelligence from commercial general liability policies. Similar exclusions are appearing in cyber, D&O, and E&O policies from many carriers. According to Insurance Journal reporting, insurer interest in AI exclusions has grown rapidly as AI use becomes more widespread.
Does the exclusion apply if my vendors use AI but I do not use it directly?
Potentially, yes. The endorsement language is broad and can apply whether the AI is used directly by the insured or indirectly through vendors, consultants, or embedded software tools. If a claim traces back to AI-driven output anywhere in the supply chain, the exclusion may apply. Reviewing vendor contracts for indemnification and insurance requirements has become more important as a result.
What is affirmative AI coverage and how do I get it?
Affirmative AI coverage is insurance specifically designed to respond to generative AI-related claims. It is available through write-back endorsements added to existing general liability, cyber, or E&O policies, or through standalone AI liability policies from carriers that have introduced them. The market is still developing, and availability varies by industry and use case. Working with an insurance advisor who is actively tracking AI coverage developments is the most efficient way to identify the right options.
How can I tell if my current policy has an AI exclusion?
Pull your policy declarations and endorsement schedule. Look for any endorsement referencing generative artificial intelligence, artificial intelligence, or machine learning. The ISO endorsement codes are CG 40 47, CG 40 48, and CG 35 08. Proprietary carrier endorsements may use different codes but similar language. If the language is not clear, ask your insurance advisor for a specific written answer about whether AI claims are covered.
Does the exclusion apply retroactively to claims from before January 2026?
No. Policy exclusions apply to the policy period in which they are added. A claim from a policy period before the exclusion was added would be evaluated under the terms of that policy. However, once the exclusion is in place at renewal, it applies to claims made during that policy period, even if the underlying AI use began earlier. This makes the renewal cycle the critical moment to review coverage.
