Skip to content

Tech E&O for healthcare AI

The specific program the Glacis MGA would create: a Tech E&O policy for healthcare AI. This page is grounded in the Law360 article (Heidi Lawson et al.), which documents the real, live exposures such a policy would cover. GROUNDED · Law360

The strategy has evolved. “The Glacis MGA underwrites it on a carrier’s paper for commission + profit share” is the Volume 1 thesis. Volume 2’s current position is that a syndicate writes and holds the book while Glacis earns on the evidence, not the underwriting — read “how Glacis underwrites it” below as the mechanics an MGA/syndicate applies to Glacis-attested risk. See the independence thesis.

A claims-made Tech E&O / cyber policy (a Casualty coverage) tailored to vendors and providers deploying AI in healthcare — sold and underwritten on a carrier’s paper via Glacis software, for commission + profit share.

The exposures it covers (grounded in real litigation)

Section titled “The exposures it covers (grounded in real litigation)”

The Law360 article documents the exact insurable events. These aren’t hypothetical — they’re in court right now: GROUNDED · Law360

  • Wrongful AI claim denial. Kisting-Leung v. Cigna Corp. (E.D. Cal., allowed to proceed March 2025): the allegation is that Cigna used AI to deny medical claims without proper review. A healthcare-AI vendor whose tool drives such denials faces professional-liability exposure. GROUNDED · Law360
  • Algorithmic discrimination. Kelly v. State Farm Fire & Casualty Co. (M.D. Ala., Oct 2025): homeowners allege “cheat and defeat AI algorithms” used as discriminatory tools that “disproportionately impact[ed] Black and non-white policyholders.” The same theory applies to healthcare AI that produces biased outcomes. GROUNDED · Law360
  • HIPAA / privacy breach. Health data is highly regulated; a breach or improper use is a first- and third-party loss (the cyber half of Tech E&O — see 06 Tech E&O).
  • Model failure / faulty output. The AI simply gets it wrong in a way that causes a provider/payer financial harm.

See 02 AI claims litigation for the case details.

Per the NAIC survey reported in the article, healthcare has the highest AI adoption (92% of health insurers report current or planned AI use)and it’s where the hottest claims litigation is (the Cigna case). High adoption + high litigation = high demand for the coverage and high need for controls. GROUNDED · Law360

How Glacis underwrites it (the whole arc, applied here)

Section titled “How Glacis underwrites it (the whole arc, applied here)”

This product is where every prior page converges:

1. The risk is NOVEL → little loss history → price with JUDGMENT RATING + write it E&S/non-admitted (pages 02, 03-entity/06) 2. Glacis evidence proves the applicant's AI governance (human-in-the-loop, bias testing, model inventory) → better RISK SELECTION + a SCHEDULE-RATING CREDIT (pages 01, 02) 3. Better-controlled risks have fewer losses → lower LOSS RATIO (page 03) → higher MGA profit share + more CAPACITY (page 04) 4. The book grows + generates data → better models + pricing → (repeat = the flywheel, section 15)

The same regulatory pressure that creates the liability (NAIC AI Model Bulletin, bias-testing expectations, third-party-vendor oversight) also creates demand for exactly the controls Glacis ships — so compliance becomes a product feature on both sides of the trade. See 14 AI insurance regulation. GROUNDED · Law360

The exposures and cases here are grounded in the Law360 article. The pricing, limits, eligibility, and program structure of any actual policy are business/ actuarial decisions to be made with the capacity carrier and counsel — not asserted here. This is a study aid, not a rate filing.

14 AI insurance regulation: NAIC AI Model Bulletin

Drill this page →6 bank questions stand behind what you just read. Check it while it’s warm.