Why Compliance Claims Are No Longer Enough—and What to Demand Instead
Healthcare organizations face an uncomfortable truth: the AI systems entering clinical workflows can claim compliance but cannot prove it. While 51% of organizations report negative AI consequences, the industry lacks methods to verify safety controls executed when decisions were made.
Download White Paper
Most security and compliance tools operate at Layer 1 or Layer 2. What they lack is Layer 3: evidence-grade attestation that third parties can independently verify.
Pre-inference filters, prompt injection defense. Vendors can catch threats.
Post-hoc analysis, dashboards. Vendors can log requests.
Cryptographic proof for third parties. No vendor can prove it.
A patient says "I had one beer at a wedding last month." The AI writes: "Patient reports daily heroin use."
| Stage | What Happened | Evidence Available |
|---|---|---|
| Spoken | "I had one beer at a wedding last month." | None retained |
| ASR Transcript | "I had one beer... heroin last month" | Possibly logged, not linked |
| LLM Processing | Interpreted as substance use disclosure | No trace of reasoning |
| Generated Note | "Patient reports daily heroin use..." | Final output only |
| EHR Write | Hallucinated diagnosis entered | Timestamp only |
The evidentiary standard healthcare organizations should demand from AI vendors before procurement approval.
Tamper-evident traces showing which controls ran, in what sequence, with pass/fail status and cryptographic timestamps.
Complete reconstruction of input context: prompts, redactions, retrieved data, and configuration state tied to each output.
Cryptographically signed, immutable receipts that third parties can validate without access to vendor internal systems.
Direct mapping to specific control objectives in ISO 42001, NIST AI RMF, and EU AI Act Article 12.
16 pages of analysis including regulatory timeline, case studies, and the complete evidence framework.
Chief Medical Officer, GLACIS Technologies
University of Washington-trained psychiatrist with extensive regulatory experience. Previously helped develop the first FDA-authorized AI diagnostic device for autism at Cognoa. She still practices clinically in Seattle and serves as courtesy teaching faculty at UW.