How can an AI tool deployment audit produce one evidence record across HTI-1, §1557, FDA PCCP, and CMS at once?
Because these regimes share a common measurement core — does the model perform as claimed, across subpopulations, and has it drifted — one reproducible evaluation can produce evidence each regime reads against its own controls. Planisphere runs the probe set…
Answer.
Because these regimes share a common measurement core — does the model perform as claimed, across subpopulations, and has it drifted — one reproducible evaluation can produce evidence each regime reads against its own controls. Planisphere runs the probe set once and emits a single sha-pinned record that maps to HTI-1's validation attributes, §1557's nondiscrimination check, a PCCP's drift bounds, and CMS's individualized-assessment framing. It is one measurement, many mappings; it does not replace each regime's separate filing or determination.
The mark behind the answer.
predictive-DSI transparency · 31 source attributes · FAVES.
More on ONC HTI-1.
Try ONC HTI-1 with a free test key.
Start with the free sandbox. Planisphere measures model behaviour and emits a reproducible, sha-pinned record — it does not certify, file, or give legal advice.