Can consequential AI risk be made underwritable?
AI is starting to help with decisions about people's health, in hospitals, and in pro sports, where teams use data about players' bodies to decide who plays and when. Darwin by Medigram is the system that checks the AI: it scores every decision on six kinds of trustworthiness and keeps a permanent record, so whether the stakes are a patient's care or a player's career, nothing happens unchecked and there's always proof. Designed to work on mobile, where the work happens.
Here is the plain version: AI risk is hard to price when an institution cannot show where consequential AI operates, who stays accountable, whether controls actually work, or what happened when a decision is questioned later. A stronger evidence architecture can make that visible instead of hidden.
A short reader path into Building Darwin, the full CEO Letter.
Why this is hard to price
The evidence an insurer needs, where consequential AI operates, who remains accountable, whether controls were tested, what happened afterward, is distributed across medicine, engineering, cybersecurity, governance, operations, vendors, and assurance. Each function can perform well while no one owns the complete evidentiary chain, and unpriced risk tends to live in exactly those interfaces.
What insurers and underwriters should ask
Can the institution identify consequential AI uses, demonstrate accountable authority, show how controls are tested, disclose material findings and remediation, and produce contemporaneous evidence after an event?
Keeping the evidentiary chain whole
The evidence an insurer needs is scattered on purpose, by design, across medicine, engineering, cybersecurity, governance, operations, vendors, and assurance. Each of those functions can do its job well while nobody owns the complete chain, which is exactly where unpriced risk tends to hide. This architecture keeps that chain whole: one path from where AI operates, to who is accountable, to whether controls were actually tested, to what the record shows afterward.
Why Sherri and the Medigram team
Sherri Douville built Darwin by working across those boundaries directly rather than handing the problem from one function to the next. She personally holds standing in the standards (IEEE/UL 2933, ANSI/HSI 2800), the architecture, the security testing, and the evidence design, so a finding in one area does not stop at a department line before it reaches the others. Medigram does not need one person to hold this together forever; the goal is to encode that same continuity into Darwin’s product and operating model, so an insurer evaluating any institution can ask for the same integrated evidence without reconstructing it after the fact.
What this does and does not establish
Better answers to these questions improve risk visibility and the quality of information available for assessment. They do not, by themselves, establish insurability, set coverage, determine pricing, or reduce losses.
Need the full architecture and evidence?
The CEO Letter contains the detailed technical rationale, verification loop, standards context and diligence path.