What insurers ask about your AI
48 questions that insurers, brokers and lawyers say come up when you buy AI liability, tech E&O or cyber cover, in 11 groups. Tick the ones you can already answer, then download the list for your broker.
These are questions you may be asked, not a standard form: no insurer we found publishes an AI application. 24 of the 48 come from an insurer or an application; each one says who raised it. Sources checked 22 October 2026.
0 of 48 questions ready. Your ticks stay in this browser only.
Slaughter and May reports underwriters moving from asking whether you use AI to asking for evidence of how you govern it.
Westfield Pro has said the key underwriting question is how you use AI, not whether you use it.
Insurers that cover AI performance price from test results: Munich Re agrees performance metrics before assessing a model, and Armilla finds AI most insurable when its error rate is known and controlled.
Human review of AI output is the control raised most often across the insurer, broker and law-firm sources we read.
Cyber insurers want to know what data goes into AI tools and who can see it.
Training data provenance and AI-made content drive intellectual property claims, and at least one application has asked about AI content directly.
What you promise customers, and what your AI suppliers promise you, decides how much liability stays with you.
Armilla’s broker page asks for details of any previous AI-related incidents.
QBE says insurers need to understand how organizations use AI and which controls matter.
New state AI laws add obligations, and with them regulatory exposure.
How much of your business depends on AI shapes the cover and limits you need.
What underwriters say they look for
- Armilla finds AI most insurable when the company can show and control a known error rate. Armilla AI, 18 Aug 2026
- Munich Re agrees performance metrics with an AI provider before its in-depth risk assessment starts. Munich Re
- Coalition says it asks companies selling AI products a lot more than a typical tech E&O account, and Zurich warns that AI risk may not be contemplated if the underwriter does not understand how a company uses AI. Zurich North America and Coalition, quoted in Business Insurance, 26 Jun 2026
- Cowbell's chief commercial officer says guardrails on AI use are the most important factor in how an insured is viewed. Cowbell, quoted in Insurance Business, 23 Sep 2026
- Westfield Specialty underwriters assess the underlying models' propensity for fraudulent claims or discriminatory output. Westfield Specialty, quoted in Risk & Insurance, 27 May 2026
- A UK broker says insurers have no consistent set of questions about how clients use AI, which leaves brokers to volunteer what they judge material. Talbot Jones, quoted in Insurance Business, 10 Aug 2026
Before you fill in an application
- Agree with your broker what counts as “AI”, including third-party tools staff use, and set a date your answers speak to. Hunton Andrews Kurth warns that incomplete or inaccurate AI answers can lead to a denied claim or a rescinded policy. Hunton Andrews Kurth, Dec 2024
- Describe your AI the same way everywhere. A broker reports submissions being held up when an application calls a company generic software while its website sells AI features. WHINS Insurance Agency, Aug 2026
- Check what your AI vendors promise you. Their indemnities shift some IP risk off you, under conditions; see the AI vendor indemnity tracker.
- Then check your current policies for AI wording with the AI clause scanner and compare who sells cover in the register.
Questions
Is there a standard AI supplemental application?
Not yet. We found no carrier or wholesale broker that publishes an AI supplemental application. One UK broker said in August 2026 that insurers have no consistent set of AI questions. The questions here come from insurers' own publications, broker checklists, law-firm guidance and one application reported by a law firm.
Does ISO/IEC 42001 or AIUC-1 certification help with insurance?
It can. Armilla says it uses ISO/IEC 42001-aligned governance evidence alongside its own technical evaluation, and that it may lead to better terms. A certificate on its own is not enough: insurers say they want ongoing evidence of testing and monitoring. See our AIUC-1 page for how that standard is tied to insurance.
What happens if I answer AI questions wrongly on an application?
A law firm that advises policyholders warns that incomplete or inaccurate answers about AI can give an insurer grounds to deny a claim or rescind the policy. Agree with your broker what counts as AI before you answer, including third-party AI tools your staff use.
Sources
- Best practices for AI disclosures in insurance applications A Philadelphia Indemnity application, as reported by Hunton Andrews Kurth, 10 December 2024
- Risk Titration AI Checklist for PI/E&O Aon, September 2026
- From ISO/IEC 42001 certification to insurable AI Armilla AI, 17 December 2025
- From paper policies to real oversight: how AI governance is becoming insurable Armilla AI, 14 January 2026
- Lessons from a year of underwriting AI liability insurance Armilla AI, 18 August 2026
- Armilla for brokers: specialized AI coverage Armilla AI, undated
- Questions insurers should be asking about their clients’ AI usage Browne Jacobson, 28 August 2025
- AI use is starting to decide which cyber risks brokers can insure Cowbell, quoted in Insurance Business, 23 September 2026
- AI and insurance: bridging innovation and risk Gallagher, 13 February 2025
- AI liability insurance: approaches to mitigating generative algorithmic risk Indemnity (UK broker), undated
- Insurance for AI companies: coverage stack for AI startups Latent Insurance, 23 February 2026
- The insurability of AI risk: a broker’s perspective Marsh, interviewed by DLA Piper, 26 June 2024
- From governance to insurance Munich Re, undated
- De-risking AI ventures: how Munich Re assesses AI performance risks Munich Re, undated
- Insuring generative AI: risks and mitigation strategies Munich Re, undated
- Insure AI (aiSure): frequently asked questions Munich Re, undated
- Understanding LLMjacking QBE, undated
- AI as a risk amplifier: how cyber insurance is evolving QBE, 17 September 2026
- Tech E&O for AI products Relm Insurance, 1 April 2026
- AI liability and model errors: where your coverage actually responds SeedPod Cyber, 30 September 2025
- How GenAI risks are reshaping the E&O insurance market Slaughter and May, 25 November 2025
- Your client’s staff are pasting data into ChatGPT. Their cyber insurer wants to know about it Talbot Jones, quoted in Insurance Business, 10 August 2026
- AI insurance requirements: insurance may not cover your AI failures Traverse Legal, 24 April 2026
- AI governance policy: business needs before the next insurance renewal Traverse Legal, 25 April 2026
- Underwriters’ dilemma: is AI a cyber or tech E&O risk? Westfield Pro, quoted in Insurance Journal, 19 March 2024
- AI risk is here, and insurers are learning to write the rules Westfield Specialty, quoted in Risk & Insurance, 27 May 2026
- How are AI exclusions showing up in technology E&O insurance? WHINS Insurance Agency, 10 August 2026
- D&O insurance: AI’s impact Woodruff Sawyer, summarized by TheCorporateCounsel.net, 24 March 2025
- AI use amplifies professional liability risks Zurich North America and Coalition, quoted in Business Insurance, 26 June 2026