Since January 1, 2026, the insurance industry’s standards body, ISO/Verisk, has offered carriers three new endorsements — CG 40 47, CG 40 48, and CG 35 08 — that let commercial general liability policies exclude claims tied to AI-generated content, and state regulators have approved more than 80% of carrier requests to add them. If your business uses AI for marketing copy, blog content, or social media and hasn’t checked its most recent renewal packet, you may already have a coverage gap you don’t know about — and if you publish AI-assisted content for a living, this is closer to your core business risk than a background IT concern.
“Does my insurance cover AI” isn’t a yes-or-no question anymore
Adoption of these endorsements isn’t universal or automatic. Verisk’s forms are optional for each carrier to attach, which means coverage is genuinely fragmented — some insurers have added the exclusion, others haven’t, and the exact wording varies by carrier and by state. The only way to know where you stand is to check your own current policy for these specific form numbers or ask your broker directly, not to assume the news coverage of this shift applies to your particular renewal.
What the new endorsements actually exclude
| Endorsement | Scope | What it excludes |
|---|---|---|
| CG 40 47 | Broad | Coverage A (bodily injury and property damage) and Coverage B (personal and advertising injury) — bars claims linked to AI-generated outputs, including defamatory content, IP infringement from AI-generated material, and physical damage traceable to an AI-driven error |
| CG 40 48 | Narrower | Only Coverage B (personal and advertising injury) — potentially preserves bodily injury or property damage coverage in limited scenarios |
| CG 35 08 | Related form | A third endorsement circulating alongside the other two — check with your broker for how it applies to your specific policy, since sourcing on its exact scope is thinner than the other two |
ISO forms underpin roughly 82% of US property and casualty policies, so once a carrier adopts one of these, it tends to spread through renewals quickly. Some carriers have gone further than the optional ISO language — one has confirmed an absolute AI exclusion across directors and officers, errors and omissions, and fiduciary lines, not just general liability.
Why this specifically matters if your business publishes AI-assisted content
CG 40 47’s own listed examples — defamatory content and IP infringement from AI-generated material — describe exactly the exposure a content-focused small business carries every time it publishes AI-assisted marketing copy, a blog post, or a social media caption. This isn’t a hypothetical connection: the new affirmative AI-liability products built to fill the gap name the same activity directly. One small-business-focused option launched in March 2026 explicitly covers losses tied to “AI-generated advertising, marketing copy, blog content, or social media” — language that could describe the daily output of a freelance writer, an agency, or a solo content operation as easily as it describes a manufacturer’s marketing department. If your business fits that description, this shift belongs in the same category of attention as verifying AI-generated content before it publishes and writing AI-assisted content that actually holds up — a real cost of doing business with these tools, not an abstract policy debate.
This isn’t the first time an insurance market has done this
The pattern matches what happened with cyber risk between roughly 2015 and 2023: for years, traditional policies “silently” covered cyber losses they were never designed to price, until insurers took enough unexpected losses to start writing explicit cyber exclusions and separate cyber policies instead. AI coverage is following the same arc, just compressed into a single year rather than spread over several — “silent AI coverage,” where a policy neither explicitly includes nor excludes AI-related claims, is what most small businesses have had by default, and 2026 is the year that default is ending for a fast-growing share of the market.
What to actually do about it
- Pull your current renewal packet and search for the specific form numbers — CG 40 47 01 26, CG 40 48 01 26, CG 35 08 01 26 — or ask your broker directly whether an AI exclusion has been added to your general liability, D&O, or E&O coverage this year.
- Treat this as a core business question if you publish AI-assisted content publicly, not a background IT or tech-risk item — the exclusion’s own examples and the new affirmative products both center on exactly that activity.
- Ask specifically about affirmative AI coverage — write-back endorsements that restore coverage a new exclusion removed, or standalone AI liability policies now being sold specifically to small and mid-size businesses — as an option to weigh against the cost of an uncovered claim.
- Check again at every renewal, not just this one. Adoption is still spreading across carriers and states through 2026, so a policy that’s silent on AI today may pick up an exclusion at its next renewal without a phone call to warn you.
This is genuinely a question for your insurance broker or an attorney familiar with your specific policy, not something a general article — including this one — can settle for your particular coverage. The uncertainty goes both ways: even where an exclusion has been added, legal commentary on these endorsements notes the broad language isn’t necessarily the final word on how a court would apply it to a specific claim.
Where this connects to decisions you’ve likely already made
If you’ve already put together a written policy for who on your team can use which AI tools — covered in our AI usage policy guide — this is a natural addition to that document’s next review: name whether your liability coverage currently excludes AI-related claims, and note the date you last checked. The same “check what you actually have, don’t assume” discipline applies here that our piece on shadow AI applies to knowing which tools are actually touching your business data. If you’re still building out your broader AI toolset, our small business AI toolkit is worth reading alongside this one rather than treating cost and coverage as separate conversations. And if EU exposure is part of your business, the disclosure obligations covered in our EU AI Act guide are a separate compliance question worth checking alongside your insurance renewal, not instead of it.
Who should treat this as urgent
Anyone whose business model involves publishing AI-assisted content that reaches the public — content creators, marketing freelancers, agencies producing client-facing copy — has the clearest, most direct exposure the new exclusions describe, and should check their current policy this month rather than at the next renewal reminder. A business using AI purely for internal drafts that a human rewrites before anything goes public carries meaningfully less of this specific risk, though it’s still worth the five minutes it takes to ask a broker the question.
Does my business insurance cover AI-related claims?
It depends entirely on your specific carrier and policy — coverage is fragmented, not universal. Check your renewal packet for endorsement forms CG 40 47, CG 40 48, or CG 35 08, or ask your broker directly rather than assuming based on general coverage of this topic.
What are CG 40 47 and CG 40 48?
They’re optional endorsements from ISO/Verisk, effective January 1, 2026, that let commercial general liability carriers exclude claims linked to AI-generated outputs. CG 40 47 is broad, covering both bodily injury/property damage and advertising injury; CG 40 48 is narrower, excluding only advertising injury claims.
Does this affect me if I only use AI for blog posts or social media, not a physical product?
Yes, potentially more directly than a physical-product business. The exclusions specifically name defamatory content and IP infringement from AI-generated material, and new affirmative AI-liability products built to fill the gap explicitly cover AI-generated advertising, marketing copy, blog content, and social media.
Is this exclusion automatic on every policy?
No. Adoption is optional per carrier and varies by state, so some businesses already have it and others don’t. The only way to know is to check your specific policy or ask your broker — don’t assume either way based on this article or any general coverage of the topic.
What’s the alternative if my policy now excludes AI-related claims?
Insurers are introducing affirmative AI-liability coverage to fill the gap — write-back endorsements that restore AI-related coverage, or standalone policies, some specifically priced for small and mid-size businesses. Ask your broker what’s available for your size and risk profile rather than assuming any one option is the right fit.
How do I find out if my policy already has this exclusion?
Search your most recent renewal documents for the specific endorsement form numbers, or ask your broker directly whether an AI exclusion has been added to your general liability, D&O, or E&O coverage. This is worth doing at every renewal for the rest of 2026, since adoption is still spreading.
Shurah is the founder of AI Tools Daily, tracking pricing, licensing and policy changes across AI tools so readers can make decisions without wading through marketing claims themselves.
