Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Short answer: Standard business insurance may cover some AI-related losses, but no policy label guarantees protection. Coverage depends on the cause of loss, the policy wording, AI exclusions, sublimits, aggregates and whether one model failure affects many insureds at once.
That uncertainty is becoming an insurance-capacity problem. Dedicated AI policies reported in 2026 generally offer limits of about $25 million to $50 million per insured, while litigation scenarios involving major AI systems can reach the billions. A single upstream model or vendor failure could also generate correlated claims across thousands of customers, overwhelming limits designed for ordinary, standalone accidents.
As an Amazon Associate I earn from qualifying purchases.
Why AI liability can outrun available insurance
The scale mismatch
The Center for Strategic and International Studies (CSIS) says dedicated AI insurance remains a boutique market, with reported limits topping out at roughly $25 million to $50 million per insured in 2026. Those limits are small beside the multibillion-dollar litigation and settlement exposures attributed to leading AI developers. A limit is also not the same as money available for one event: deductibles, retentions, defense costs, sublimits and aggregate caps can reduce what an insured actually collects.
One failure can create many claims
Traditional underwriting often assumes that losses are reasonably independent. Foundation models challenge that assumption. If one upstream model produces unsafe instructions, discriminatory decisions, a misleading answer or infringing material, the same defect can reach thousands of downstream products and customers. Insurers therefore worry about accumulation: many policies responding to one technical or vendor event at the same time.
AI errors do not fit neatly into one coverage line
An erroneous output might be an operational mistake, a professional-service failure, defamation, misrepresentation, an intellectual-property dispute, a privacy incident or a physical-safety event. Cyber insurance is usually built around security incidents and data breaches, so a wrong chatbot answer may not satisfy its insuring agreement. Commercial general liability (CGL) wording may address bodily injury, property damage or personal and advertising injury, but an AI endorsement can remove those protections.
What the 2026 market signals mean
Gartner advises general counsel to evaluate “affirmative AI insurance” for hallucinations, bias, intellectual-property infringement and safety failures. Gartner also forecasts more than 2,000 “Death by AI” legal claims worldwide by the end of 2026. That is a forecast, not a tally of cases already filed.
“GC should be evaluating a new wave of ‘affirmative AI insurance’ offerings that provide targeted coverage for risks, such as hallucinations, bias, IP infringement and safety failures.” — Gartner, 2026
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteSpecial offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Econofact reports that major AI companies have struggled to buy enough emerging-risk insurance for potential litigation costs. Existing cyber and CGL policies may respond to some AI losses, but only according to their wording. Aon similarly says AI exposures are embedded across policies that were not designed for the technology, leaving uncertainty over whether a claim is covered at all.
The uncertainty is reflected in the market’s own description. Thomas Bentz, a Holland & Knight partner, told the IAPP in 2026: “I think there’s a lot of confusion and growth in this industry, because where these claims fit is still kind of being figured out.”
Compare the insurance lines, not just the policy name
| Coverage line | Traditional focus | Typical AI coverage question |
|---|---|---|
| Commercial general liability (CGL) | Bodily injury, property damage, and personal and advertising injury | Does an AI exclusion remove injury, damage, defamation or advertising claims? |
| Cyber | Security incidents and data breaches | Does an erroneous or misleading output trigger a covered security event, or is it outside the insuring agreement? |
| Technology E&O/professional liability | Failures in technology services and other professional services | Does the wording cover probabilistic model behavior, inaccurate advice, service failure and defense costs? |
| D&O and fiduciary liability | Claims against directors, officers and fiduciaries | Could allegations about AI oversight, governance or disclosure reach the policy, and are AI exclusions present? |
| Products/completed operations | Injury or damage caused by products or completed work | Is AI embedded in a product or operational system, and do products/completed-operations exclusions apply? |
Ropes & Gray notes that an AI-generated error may produce defamation or misrepresentation allegations rather than the security incident or personal-data breach required by many cyber forms. Insurers are also adding AI carve-outs, exclusions and AI-specific sublimits.
What the 2026 ISO forms change
The reproduced ISO endorsement CG 40 47 01 26 states that CGL insurance does not apply to “bodily injury” or “property damage” arising out of “generative artificial intelligence.” It separately excludes “personal and advertising injury” arising out of generative AI.
The endorsement defines the term as follows:
“Generative artificial intelligence” means a machine-based learning system or model that is trained on data with the ability to create content or responses, including but not limited to text, images, audio, video or code.
Related 2026 ISO forms address narrower coverage parts, including Coverage B-only and products/completed-operations exclusions. These forms are optional. Their effect depends on whether a carrier attaches them to a particular policy and how the relevant state filing is handled. The existence of a form therefore does not prove that every CGL policy excludes AI; the attached endorsements and declarations control.
How common AI incidents may be treated
A chatbot makes a defamatory statement
The alleged harm may look like personal and advertising injury under CGL, or a professional-service error under technology E&O. A generative-AI exclusion, an AI sublimit or a requirement that the statement result from a specified offense can defeat coverage. Cyber insurance is not automatically triggered because no security incident or data breach occurred.
An AI system contributes to bodily injury
Product liability, products/completed operations and CGL are the lines most likely to be examined when an AI-enabled machine, vehicle, device or operational system causes physical harm. The 2026 ISO wording shows how a carrier can expressly exclude generative-AI bodily injury and property damage, so the policy must be checked for that endorsement before relying on the coverage.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →An output infringes intellectual property
Potential theories include advertising injury, professional negligence or a technology-service failure. Gartner identifies IP infringement as a target for affirmative AI insurance because conventional forms may contain intellectual-property exclusions, narrow definitions of advertising injury or no coverage for the insured’s own development activity.
An AI decision causes discrimination or misleading advice
Bias, hallucinated advice and inaccurate automated decisions can create regulatory, contract, professional-liability and governance allegations. The applicable response may involve E&O, D&O or a specially written AI policy rather than cyber insurance. Coverage depends on the alleged act, the insured entity, the jurisdiction and the policy’s defense and indemnity wording.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Pre-renewal AI coverage audit
Companies can reduce surprises by completing a documented review before renewal or deployment:
- Inventory systems and vendors. Record every foundation model, application, plug-in, data source and business process that uses AI, including systems operated by suppliers.
- Map plausible harm. For each system, identify hallucinations, bias, IP infringement, privacy exposure, cyber compromise, unsafe operation, bodily injury, property damage and business-interruption scenarios.
- Read the definitions. Search every policy for “artificial intelligence,” “generative artificial intelligence,” “autonomous system,” “AI-assisted decision” and related terms. Check whether the definition covers development, deployment, use and outputs.
- Locate exclusions and sublimits. Mark AI exclusions, technology exclusions, intellectual-property exclusions, professional-services exclusions, retentions, aggregate limits and systemic-risk language.
- Test trigger language. Ask what event starts coverage: a security incident, breach, error or omission, publication of content, bodily injury, property damage, claim, occurrence or wrongful act.
- Check coordination. Compare CGL, cyber, technology E&O, D&O, fiduciary, product-liability and completed-operations policies for overlapping grants, gaps and conflicting exclusions.
- Prepare underwriting evidence. Keep records of model testing, human review, access controls, vendor contracts, incident escalation and board-level governance. Underwriters may require this evidence before offering affirmative coverage or higher limits.
Questions to put to the broker and carrier
- Which AI systems, vendors and outputs are covered, and which are excluded?
- Does the policy cover development, deployment, customer use and third-party models separately?
- Are AI losses subject to a sublimit, a separate retention, an aggregate cap or a systemic-loss exclusion?
- How are multiple claims from one model, vendor or software release aggregated?
- Are defense costs inside or outside the limit, and when must an incident be reported?
- Which jurisdictions are covered, and do regulatory investigations or class actions qualify as claims?
- Do governance warranties require human oversight, testing, documentation or specific controls?
- Can the broker obtain affirmative AI coverage, excess limits, a captive arrangement or another risk-transfer structure if the primary program is insufficient?
There is no universal AI-liability policy yet. The practical decision is whether the company can identify a covered trigger and enough limit for its most severe, correlated scenario. If it cannot, the gap should be documented and addressed through wording changes, additional capacity or operational controls rather than assumed away under a familiar policy label.
Recommended Free Tools
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
