AI Liability Insurance Buyer's Guide

AI liability insurance carrier comparison

By Joel Singh · Sources monitored daily, last checked 2026-10-01 · full source-by-source review by a person quarterly, last 2026-07-18

Every US AI liability insurance product we track: standalone policies, performance guarantees, DIC wraps, and endorsements. Every cell that states a fact links to its source; products pending primary-source verification are flagged. The underlying dataset is at /data/carriers.json.

7 Standalone policy3 Endorsement1 Performance guarantee1 DIC wrap

Disclaimer: This comparison is informational only. It is not insurance advice, legal advice, or a recommendation of any policy. Coverage terms, limits, and availability change frequently and vary by state and by insured. Always verify current terms directly with the carrier or a licensed broker before making any coverage decision.

Quick reference

Click a column header to sort. Filter chips below also hide matching rows.

Product ▴▾ Category ▴▾ Limits ▴▾ Segment ▴▾ How to buy ▴▾ Geography ▴▾ Verification ▴▾
Armilla Affirmative AI Liability / Vanguard AI Standalone AI liability $25M+ dedicated AI aggregate limits alongside $10M cyber (Vanguard AI structure) [src] Enterprise AI vendors and deployers [src] Surplus lines broker; Armilla requires an AI system assessment/certification before writing a policy [src] Not disclosed / verify with carrier [src] checked 2026-10-01
Munich Re aiSure (direct and via Mosaic Insurance) AI performance guarantee Up to USD/EUR/CAD 15M initial capacity via Mosaic [src] AI developers and vendors who need to backstop contractual performance promises [src] Broker to Mosaic or Munich Re [src] Worldwide (Mosaic capacity for AI developers and vendors) [src] checked 2026-10-01
Testudo GenAI Liability Standalone AI liability $10M per insured, per Testudo, direct confirmation, August 2026; capacity was $9.25M in early 2026 [src] Any U.S. company deploying generative AI in its products or operations [src] Broker placement via Testudo (Lloyd's coverholder); distributed on an excess and surplus lines (E&S) basis in the US, through your existing brokers [src] US focus (mid-market) [src] checked 2026-10-01
HSB AI Liability Insurance for Small Businesses Standalone AI liability (SMB, distributed via partner carriers) Not disclosed / verify with carrier Small and mid-sized businesses (SMB) [src] Not sold direct. Added to business policies of carriers that partner with HSB, pending state regulatory approval [src] US, subject to state regulatory approval [src] checked 2026-10-01
AIUC AI Agent Insurance (AIUC-1 certified) Vendor-bundled AI agent insurance with certification standard Not disclosed / verify with carrier Enterprises deploying AI agents; AI vendors bundling insurance with their product [src] Via AIUC or bundled by participating AI vendors (for example ElevenLabs ElevenAgents); AIUC-1 certification required before insuring an agent [src] Not disclosed / verify with carrier checked 2026-10-01
Vouch AI Insurance (with Corix, a Hiscox division) Cyber/E&O endorsement tier (startup-focused AI coverage) Not disclosed / verify with carrier VC-backed AI startups [src] Direct via Vouch (digital broker/distributor) [src] US checked 2026-10-01
Relm NOVAAI Cyber + tech E&O for AI companies Not disclosed / verify with carrier AI platform and product companies [src] Broker to Relm (Bermuda specialty market) Bermuda specialty market; verify admitted/surplus status for US insureds checked 2026-10-01
Relm PONTAAI Excess DIC wrap for AI exclusions Not disclosed / verify with carrier AI deployers and developers whose existing programs exclude AI [src] Broker to Relm (Bermuda specialty market) Bermuda specialty market; verify admitted/surplus status for US insureds checked 2026-10-01
Relm RESCAAI First-party AI loss coverage Not disclosed / verify with carrier Companies exposed to first-party AI losses Broker to Relm (Bermuda specialty market) Bermuda specialty market; verify admitted/surplus status for US insureds checked 2026-10-01
Coalition Affirmative AI Endorsement Cyber endorsement (AI as attack vector, not AI output liability) Not disclosed / verify with carrier Cyber policyholders (US surplus and Canada) Via Coalition cyber policy (broker-distributed) US surplus lines and Canada [src] checked 2026-10-01
Embroker AI Coverage Endorsement Tech E&O / cyber endorsement (automatic on eligible quotes) Not disclosed / verify with carrier Startups (Embroker Startup Program) Direct via Embroker digital platform; automatic on eligible quotes [src] US checked 2026-10-01
CFC Affirmative AI Coverage Affirmative AI embedded across core commercial lines Not disclosed / verify with carrier Not disclosed / verify with carrier Broker to CFC Not disclosed / verify with carrier checked 2026-10-01

Product details and sources

Full coverage summaries, covered perils, notable exclusions, and the source behind each verified cell. Click any row to expand.

Armilla Affirmative AI Liability / Vanguard AI

Standalone policy checked 2026-10-01

Underwriter: Chaucer and other Lloyd's underwriters (Armilla as Lloyd's coverholder) · Launched: April 2025 (affirmative AI liability); February 2026 (Vanguard AI coordinated structure) [src]

First affirmative AI liability policy. Triggered by AI underperformance: hallucinations, model drift, mechanical failures, and critical errors leading to damages. Includes legal defence and liability. Vanguard AI pairs dedicated AI aggregate limits with cyber limits under predefined allocation rules for mixed cyber/AI scenarios. A separate Armilla Guaranteed product provides a performance warranty paying compensation if the AI fails contractual KPIs such as accuracy or bias thresholds. [src]

What it covers

  • Hallucinations and critical AI errors causing damages
  • Model drift
  • Mechanical failures of AI systems
  • Legal defence costs and liability
  • Performance warranty option (Armilla Guaranteed) for contractual KPI failures

Notable exclusions

Not disclosed / verify with carrier

Sources

Munich Re aiSure (direct and via Mosaic Insurance)

Performance guarantee checked 2026-10-01

Underwriter: Munich Re; distributed and underwritten via Mosaic Insurance cyber specialists since February 2026 · Launched: 2018 (Munich Re aiSure); February 2026 (Mosaic partnership) [src]

AI performance guarantee insurance. No negligence allegation needed: breach of a predefined performance threshold triggers payout, similar to a parametric structure. Covers financial loss from defined AI performance failures, with claims settled on measurable performance data. Underwriting focuses on the model itself rather than the insured's industry. [src]

What it covers

  • Financial loss from defined AI performance failures
  • Breach of predefined performance thresholds (no negligence allegation required)
  • Backstop for contractual performance promises made by AI vendors

Notable exclusions

Not disclosed / verify with carrier

Sources

Testudo GenAI Liability

Standalone policy checked 2026-10-01

Underwriter: Lloyd's panel including Apollo, Atrium, and QBE (Testudo as MGA, Lloyd's Lab-backed) · Launched: Late 2025 (platform announced Jun 2025, Lloyd's Lab-backed; capacity expanded to $9.25M Mar 2026, then increased to $10M per insured August 2026) [src]

Generative AI liability insurance for vendors and deployers of AI. Covers third-party claims arising from AI-generated outputs, including hallucinations, model drift, and deviations from expected behavior, alongside IP infringement, unauthorized data disclosure, bodily injury and property damage, and regulatory proceedings. Includes legal costs and damages in scenarios where CGL may not respond. Underwriting uses a proprietary platform mining AI litigation, regulatory, and incident data. [src]

What it covers

  • Third-party claims from AI-generated outputs
  • Hallucinations
  • Model drift and deviations from expected behavior
  • IP infringement
  • Personal injury
  • Unauthorized data disclosure
  • Bodily injury and property damage
  • Regulatory proceedings (AI regulatory violation)
  • Legal costs and damages where CGL may not respond

Notable exclusions

Not disclosed / verify with carrier

Sources

HSB AI Liability Insurance for Small Businesses

Standalone policy checked 2026-10-01

Underwriter: HSB (a Munich Re subsidiary) · Launched: Announced 2026-03-18 [src]

First standalone AI liability product for small and mid-sized businesses. Covers lawsuits arising from the use of AI, including bodily injury, property damage, and advertising injury from AI-generated advertising, marketing, blogs, and social media. Designed to fill gaps that general liability policies exclude. [src]

What it covers

  • Lawsuits from use of AI
  • Bodily injury and property damage arising from AI use
  • Advertising injury from AI-generated advertising, marketing, blogs, and social media

Notable exclusions

Not disclosed / verify with carrier

Sources

AIUC AI Agent Insurance (AIUC-1 certified)

Standalone policy checked 2026-10-01

Underwriter: The Artificial Intelligence Underwriting Company (AIUC) · Launched: Launched with $15M funding; first live deployment February 2026 (ElevenLabs voice agents) [src]

Combines the AIUC-1 certification standard (5,000+ adversarial simulations across security, safety, reliability, privacy, and accountability) with insurance for AI agents. ElevenLabs became the first to go live with AIUC-1-backed agent insurance for voice agents deployed via ElevenAgents, an early example of insurance bundled by the AI vendor. [src]

What it covers

  • Insurance for AI agents backed by AIUC-1 certification
  • Certification spans security, safety, reliability, privacy, and accountability testing

Notable exclusions

Not disclosed / verify with carrier

Sources

Vouch AI Insurance (with Corix, a Hiscox division)

Endorsement checked 2026-10-01

Underwriter: Corix Insurance Services LLC and Vouch Insurance Company (both acquired by Hiscox, announced August 2025, subject to regulatory approval); Vouch Inc. remains an independent broker under a multi-year distribution agreement with Hiscox · Launched: Vouch AI Insurance line for startups; Corix sold to Hiscox August 2025

AI insurance for startups: AI E&O covering hallucinations and other harmful or misleading model outputs, algorithmic bias and discrimination, regulatory investigation defence costs, and IP infringement claims tied to training data or generative outputs. Offered through the Corix (Hiscox) partnership. [src]

What it covers

  • AI E&O (hallucinations)
  • Algorithmic bias and discrimination
  • Regulatory investigation defence costs
  • IP infringement claims from AI systems

Notable exclusions

Not disclosed / verify with carrier

Sources

Relm NOVAAI

Standalone policy checked 2026-10-01

Underwriter: Relm Insurance (Bermuda specialty) · Launched: January 2025 [src]

Cyber and technology E&O coverage built for AI platform and product companies, one of three AI liability solutions Relm launched in January 2025. [src]

What it covers

  • Cyber coverage for AI platform/product companies
  • Technology E&O for AI platform/product companies

Notable exclusions

Not disclosed / verify with carrier

Sources

Relm PONTAAI

DIC wrap checked 2026-10-01

Underwriter: Relm Insurance (Bermuda specialty) · Launched: January 2025 [src]

Excess difference-in-conditions (DIC) wrap for deployers whose existing insurance programs exclude AI. Covers professional negligence, IP infringement, personal injury, discrimination, privacy, bodily injury and property damage, and civil fines (where legally insurable) for violations of AI-specific regulations such as the EU AI Act and the Colorado AI Act. The most direct market answer to the 2026 AI exclusion wave. [src]

What it covers

  • Professional negligence
  • IP infringement
  • Personal injury
  • Discrimination
  • Privacy
  • Bodily injury and property damage
  • AI regulation violations

Notable exclusions

Not disclosed / verify with carrier

Sources

Relm RESCAAI

Standalone policy checked 2026-10-01

Underwriter: Relm Insurance (Bermuda specialty) · Launched: January 2025 [src]

First-party coverage for AI-related losses: business interruption, reputational harm, product recall, and incident response. [src]

What it covers

  • Business interruption
  • Reputational harm
  • Product recall
  • Incident response

Notable exclusions

Not disclosed / verify with carrier

Sources

Coalition Affirmative AI Endorsement

Endorsement checked 2026-10-01

Underwriter: Coalition (cyber MGA) · Launched: Affirmative AI endorsement announced 2025; deepfake reputational-harm coverage December 2025

Affirmative AI endorsement on US surplus and Canada cyber policies. Expands the security failure and data breach definition to include AI security events. December 2025 addition: deepfake reputational-harm coverage. Important scope distinction: this covers AI as an attack vector, not liability for your own AI's outputs. [src]

What it covers

  • AI security events (expanded security failure / data breach definition)
  • Deepfake reputational-harm coverage (December 2025)

Notable exclusions

Does not cover AI output liability; scope limited to AI as attack vector [src]

Sources

Embroker AI Coverage Endorsement

Endorsement checked 2026-10-01

Underwriter: Embroker (digital broker); underwriting carrier not disclosed in cited source · Launched: 2025-08-05 [src]

AI Coverage Endorsement automatically included on eligible tech E&O and cyber quotes in the Embroker Startup Program since August 5, 2025. [src]

What it covers

  • AI coverage endorsement on eligible tech E&O quotes
  • AI coverage endorsement on eligible cyber quotes

Notable exclusions

Not disclosed / verify with carrier

Sources

CFC Affirmative AI Coverage

Standalone policy checked 2026-10-01

Underwriter: CFC · Launched: Not disclosed / verify with carrier

CFC embeds affirmative AI coverage across seven core products (technology E&O, professional E&O, eHealth, IP, management liability, media, and cyber CPR) as of June 2026. Policies address AI-specific exposures including hallucinations, AI-generated content, and model drift. A July 2026 extension adds affirmative AI to the media policy. [src]

What it covers

  • Affirmative AI cover embedded across seven core lines
  • Hallucinations
  • AI-generated content
  • Model drift
  • Media policy (affirmative AI extension, July 2026)

Notable exclusions

Not disclosed / verify with carrier

Sources

Changelog

Every change to this dataset is dated and described. Cells are re-verified against their sources at least quarterly; launches and capacity changes trigger ad hoc updates. See the editorial policy.

Free · No spam · Unsubscribe anytime

Get the AI Coverage Buyer's Brief

Subscribe to get the Buyer's Brief now, plus the monthly update: exclusion-wave changes, new carriers, verification diffs. Unsubscribe anytime.

Editorial independence protected. Rankings and verification are never for sale.

HOME