Clinical AI risk is already inside your MPL treaties.
Nobody has priced it, and MPL underwriting was not built to. Idem is building the specialist that assesses and insures clinical AI, and we are looking for partners to back it.
Silent AI.
MPL rates rest on loss history that predates clinical AI. The exposure is covered but not rated, and it flows through to the treaties behind those policies. Other lines are closing the door: ISO generative AI exclusions took effect in general liability in January 2026, and W.R. Berkley filed an absolute AI exclusion on D&O and E&O.
Meanwhile the mistakes are real. When Ontario's Auditor General tested the 20 AI scribes approved for clinical use, every one produced errors in the medical record.
The reinsurers that backed specialists came through best.
Cyber moved from silent cover to exclusions to a loss shock, then became its own line priced on each insured's controls. The capital that did well backed specialists that assess and insure in one loop: Allianz with Coalition, Munich Re with At-Bay, which it agreed to buy for $575M in August 2026. US cyber premium grew 3.8 times from 2017 to 2023.
Clinical AI is at the exclusion stage now. Unlike cyber, it cannot be scanned from outside, and malpractice claims take years to develop, so the assessment data has to exist before the losses do.
The assessment, the insureds, and the data.
A record of which AI tools each insured runs, so AI involvement can be traced when claims arrive.
Exposure mapped by vendor and model. One vendor update can reach hundreds of insureds at once, so aggregates and event definitions need this map.
Each renewal pairs an assessment with a claims outcome, which is how we test whether our tiers predict losses.
A published standard, and no vendor pays for an insured's result.
Partners to fund and back the program.
We are looking for carrier and reinsurance partners to back a clinical AI coverage program, written on MPL or fronting paper and priced on the Idem tier. The coverage is built for hallucinations, wrong or missing outputs, model drift after vendor updates, and uneven performance across patient groups. We want partners for co-design of the form, tiers, and event definition, capacity for a pilot, and a shared view of assessment and claims data as it builds.
Treaty and facultative underwriters writing medical professional liability, casualty and emerging-risk actuaries, accumulation and cat modeling teams, and reinsurance brokers placing healthcare programs.
Read the evidence.
Then let's talk.
The June 2026 white paper documents what EHR-native AI features are doing in the medical record and what your organization is on the hook for.