AI should support clinical judgment without making accountability ambiguous.
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.
For medicine, trustworthy AI is not just a model-performance question. The system must make clear what AI was allowed to contribute, who remained clinically accountable, when escalation was required, and what evidence exists if the decision is questioned later.
A short reader path into Building Darwin, the full CEO Letter.
Why medicine has a different trust threshold
Clinical work can produce irreversible consequences. TTIC’s origin story argues that medicine already has established trust machinery in accreditation and published literature: trust depends on how a conclusion was produced, not only on the conclusion itself.
What Darwin changes for the physician
The clinical recommendation is only one field inside a governed decision record. The architecture also captures the model, policy, governance context, security controls, verification and evidence supporting the decision.
Infrastructure should absorb the seams, not the physician
When AI-related responsibilities are split across technology vendors, IT, security, compliance, and administration, the physician at the point of care is often the one left to notice when something doesn’t add up, and to absorb the consequences if it doesn’t. This architecture is built so that governance of those seams is deliberate and upstream, not something clinicians have to compensate for in the moment.
Why Sherri and the Medigram team
That design reflects work done across standards, architecture, engineering, security, and evidence together, so a finding in one area (a testing result, a boundary condition) changed decisions in the others without waiting on a handoff. The intent is for that discipline to live in the infrastructure itself, so physicians can rely on defined boundaries and preserved evidence rather than becoming the integration layer for gaps another system left open.
What happens when inputs are insufficient
Darwin is designed to preserve the governance record even when the inputs do not support a recommendation. The result is visible evidence of uncertainty or validation failure rather than silence.
The practical promise
If a decision is reviewed three days or three years later, the physician should not have to reconstruct the situation from memory, screenshots and email. The contemporaneous record should already exist.
Need the full architecture and evidence?
The CEO Letter contains the detailed technical rationale, verification loop, standards context and diligence path.