Does Your Business Insurance Cover AI? What Changed in 2026

Your business probably uses AI already — a chatbot on your website, an assistant drafting emails, a tool that helps with scheduling, bookkeeping, or customer service. Most owners assume their existing insurance has them covered. Through 2026, that assumption started to break.

Insurers spent the year doing something new: writing AI into their policies on purpose — and often writing it out. Coverage that used to apply quietly (the industry calls it "silent AI") is being replaced with explicit language that either covers AI or excludes it. Which one you got depends on your carrier and your last renewal.

This guide explains, in plain English, what your policies actually cover when AI is in the mix — and the handful of things worth checking before your next renewal.

The coverages AI touches

AI doesn't create one new risk. It touches several policies at once:

• 1. Cyber liability — when AI is the weapon in an attack, or the leak in a data breach.
• 2. Professional liability (E&O) — when your AI does the work and gets it wrong.
• 3. General liability — the policy most likely to have added a new AI exclusion.
• 4. Media and management liability — for AI-generated content, and for how leadership governs AI.

The mistake is assuming one policy covers all of it. None of them does.

What Changed in 2026

For years, AI coverage was "silent" — policies didn't mention AI, so a claim involving it usually fell under whatever coverage fit best. That era is ending. In 2026, carriers started making AI explicit, and the language cuts both ways.

• In early 2026, the industry's main policy-form organization introduced optional endorsements that let carriers exclude generative-AI claims from general liability — including bodily injury, property damage, and advertising injury.
• Professional and management lines (E&O, directors & officers, employment practices) are the least consistent. Some carriers added outright AI exclusions; others offer an affirmative "buy-back" that puts the coverage back for a price.
• Cyber carriers mostly kept covering AI-driven attacks, but some added sublimits that cap what they'll pay on an AI-related loss.
• The first standalone AI liability products arrived for small and mid-size businesses.

The takeaway: AI coverage is now something you negotiate, not something you assume. Two businesses doing identical work can end up with opposite coverage, depending on which endorsement landed on their renewal.

Does Your Cyber Policy Cover AI?

This is the question we get most, and the honest answer is: partly. Cyber liability covers AI as a weapon or a leak — not as a bad employee.

Here's the line that sorts it out. If the AI breaks in or spills data, that's cyber. If the AI does the work badly, that's not.

Cyber usually responds when:

• An attacker uses AI against you — a deepfake voice, AI-written phishing, AI-enhanced social engineering that tricks your team into a wire transfer.
• A tool you use leaks data — an employee pastes client information into an AI assistant and it's exposed, or an AI vendor you rely on gets breached and your data goes with it.
• AI is simply part of a normal security event — ransomware, a network outage, business interruption. The cause doesn't change the coverage.

Cyber usually does not respond when:

• The AI is just wrong — it hallucinates, gives a client bad output, or your chatbot promises something you can't honor. That's a performance failure, not a security breach.
• AI-generated content infringes someone's copyright, or an AI hiring tool screens out a protected group. Those are different exposures entirely (more below).

And even inside cyber's lane, watch for that 2026 sublimit. Your policy might cover an AI-driven attack — but only up to a capped amount that's smaller than your main limit. Read the endorsement, or have someone read it for you.

One subtlety worth knowing: some policies define a "breach" as an unauthorized disclosure, outside management's knowledge. An employee quietly pasting client data into a public AI tool can fit that definition — while a tool your company officially approved mishandling the same data might not, because you authorized the tool. It's counterintuitive, but it's how some forms read, and it's one more reason to check the definition rather than trust the headline.

Where AI Claims Actually Land

Because AI touches several policies, the real question isn't "am I covered?" It's "which policy covers this kind of AI problem?" Here's the map:

If AI causes this	The coverage that responds
A tool is breached and leaks client data	Cyber liability
A deepfake or AI phishing attack triggers a fraudulent transfer	Cyber liability / crime
Your AI gives a client bad work that costs them money	Professional liability (E&O)
AI-generated content infringes a copyright	Media liability
An AI tool you use in your own hiring screens out a protected group	Employment practices (EPL)
An AI decision leads to bodily injury or property damage	General liability (if not excluded)
You build or resell an AI product that fails	Technology E&O

One nuance the table hides: employment practices (EPL) covers your own hiring. If you build or sell an AI screening tool and a bias claim comes from your customer's applicants, that's a different exposure — and many technology E&O and cyber forms exclude discrimination claims outright. If you make AI hiring tools, treat that as a gap to confirm, not assume.

Professional liability (E&O) is where most AI claims go

If AI helps do your work — and you're a consultant, agency, accountant, firm, or any service business — professional liability is the policy that matters most. When the AI produces a flawed deliverable and a client suffers a financial loss, that's a classic errors-and-omissions claim. The catch in 2026 is that some E&O carriers are narrowing their AI language, so the coverage you count on may need a buy-back to stay intact. This is the single most important policy to check if AI is anywhere in how you serve clients.

General liability — and the new AI carve-out

General liability covers bodily injury and property damage, so most AI claims (which are financial, not physical) never touched it anyway. What changed in 2026 is the new optional exclusion carriers can bolt on. If your business has any real-world AI exposure — robotics, automated equipment, AI-driven physical systems — confirm that exclusion isn't on your policy.

Content, hiring, and the boardroom

Three more exposures round it out. AI-generated content that copies a protected work is a media and intellectual-property issue, not a cyber or GL one. AI in hiring — resume screeners, automated candidate ranking — can create discrimination claims that land under employment practices liability. And for larger or funded companies, "AI washing" (overstating your AI capabilities to investors or customers) is a governance exposure that points back at directors & officers coverage — and it doesn't stop there. Overstating what an AI product actually does can also trip the deceptive-practices and "failure to conform to represented performance" exclusions in technology E&O, which is how a marketing claim quietly becomes an uninsured one.

A Few AI Risks That Deserve Their Own Look

Three AI exposures are big enough — and misunderstood enough — that we wrote a separate guide on each:

• Wiretap lawsuits from chatbots and note-takers. Your website chat tool or AI meeting recorder can trigger a recording-law claim, and many cyber and E&O forms exclude exactly that. See AI chatbots, note-takers, and wiretap lawsuits.
• Deepfake wire fraud. Social engineering coverage can pay, but a sublimit, a retention, and a callback condition decide how much. See what your social engineering coverage actually pays.
• An AI or cloud vendor going down. Lost income from a provider outage falls under dependent business interruption — capped, delayed by a waiting period, and sometimes excluded outright. See when your AI vendor goes down.

What Washington Businesses Should Check

You don't need to become an insurance expert. You need to ask a few specific questions at renewal:

• 1. Ask whether your renewal added an AI exclusion. Carriers aren't required to flag it. Have your agent confirm, in writing, whether generative AI is excluded on your general liability and professional liability.
• 2. Match your AI use to the right policy. If AI helps deliver your work, your E&O is the one to protect. If AI mostly touches data and systems, focus on cyber.
• 3. Look for an AI sublimit on your cyber policy. A capped limit isn't the same as full coverage. Know the number.
• 4. Ask about technology E&O if you build or resell AI. Standard professional liability may not contemplate selling an AI product that fails. If you're a tech or software company, our technology & cyber insurance program combines tech E&O and cyber on a single form.
• 5. Don't rely on "silent" coverage. The gaps carriers are writing in are specific and intentional. Assuming you're covered is exactly how businesses get surprised at claim time.
• 6. Answer the AI questions on your renewal truthfully. Applications now ask whether and how you use AI, and your answers become part of the policy. A wrong answer can sink a claim later. If your AI use changed this year, say so — especially heading into December renewals.

The practical move is a short policy review. We'll read the exclusions against how your business actually uses AI and tell you, plainly, where you stand. Start a review here and we'll walk through it together.

What AI Insurance Usually Won't Cover

Even with the right policies in place, some AI-related costs typically fall outside coverage. Know these going in:

• Patent claims — allegations that your AI or its output infringes a patent.
• Training-data copyright — claims over the data a model was trained on.
• Fixing or retraining the model — the cost to repair or rebuild an AI system after it fails.
• Contractual AI promises — indemnities you give a customer in a contract, which your policy never agreed to.
• An AI agent that sends money by mistake — a loss that's neither fraud nor a breach, so neither crime nor cyber may respond.

These gaps are why the contracts you sign and the way you deploy AI matter as much as the policy you buy. A good review covers both.

Frequently Asked Questions

Does cyber insurance cover AI?

Partly. Cyber liability covers AI as a weapon or a leak — an AI-driven attack against you, or a data breach through an AI tool. It does not cover the AI simply being wrong: a bad output, a hallucination, or a flawed deliverable. Those are professional liability (E&O) claims. In 2026, some cyber carriers also added sublimits that cap what they'll pay on an AI-related loss, so check your endorsement.

Does general liability cover AI?

Rarely, and less than it used to. General liability covers bodily injury and property damage, which most AI claims aren't. On top of that, carriers can now add an optional endorsement that excludes generative-AI claims from GL. If your business has physical AI exposure, confirm that exclusion isn't on your policy.

What insurance covers AI mistakes?

When AI does your work and gets it wrong, professional liability (errors and omissions) is the coverage that responds. That's true whether a person or a tool made the error. The risk in 2026 is that some E&O policies are adding AI exclusions or requiring a buy-back, so the coverage needs to be confirmed rather than assumed.

Do small businesses using ChatGPT need special insurance?

Not necessarily a separate policy, but they do need to check the ones they have. Even light AI use creates exposure — a leaked prompt, a bad AI-written answer sent to a client, an AI-generated image that infringes a copyright. The goal is making sure your existing cyber and professional liability respond, and that no new exclusion quietly removed them.

Are insurers really excluding AI in 2026?

Yes, though not universally. The industry moved from "silent" AI coverage to explicit language this year. Some carriers exclude AI, some offer affirmative buy-backs, and new AI-specific products are entering the market. The result is that coverage varies a lot by carrier — which is exactly why a policy review is worth doing.

Related Reading

• AI and Professional Liability: Is Your E&O Still Covering You? — the work-product side of AI risk, for service businesses
• Your Employees Are Using AI. Does Your Cyber Policy Cover It? — the data-leak and "shadow AI" exposure
• Is Cyber Liability Insurance Worth It? — what cyber actually covers, and when it earns its premium
• Understanding Professional Liability Insurance — how E&O responds when the work goes wrong
• Cyber Liability Coverage — the policy that guards your doors and your data