Introducing the DyFram R Score A Single Number Risk Thermometer for AI Systems
Subtitle The DyFram R score replaces vague risk adjectives with a precise auditable metric derived from the DyFram Classification Signature DCSig
What the R score is
DyFram R score is a single number risk thermometer for AI systems. Where governance debates long relied on vague labels such as high risk or low risk the R score replaces adjectives with a precise auditable metric derived from a systems DyFram Classification Signature DCSig. The R score lets regulators procurers and operators say not that something feels risky but that it reads at a specific reproducible value.
How it is produced
The R value is computed from the Multi Dimensional Categorization Matrix MDCM. MDCM captures a systems capability profile use case actor classification ownership class impact profile control mode and context. Those seven dimensions form a deterministic DCSig. DyFram applies a transparent weighting and normalization function to the DCSig to produce R a numeric representation of the systems aggregated risk posture. Every step is versioned and recorded so the R calculation is auditable.
Why a single number matters
A single comparable value removes semantic drift. Governments can define R ranges that map directly to standardized G level s; those G level s trigger layered governance bundles UGC CSG UGL OGL and conditional paths CPP MCR. That mapping preserves both local policy nuance through JAL and cross deployment comparability because the meaning of G level s remains fixed while jurisdictions tune sensitivity.
Governance mechanics made operational
R is intentionally simple to read and rigorous under the hood. It is not a policy decision in itself it is an operational input. An R reading becomes a deterministic governance anchor classification provenance DCSig explains why a system scored as it did the standardized G level tied to that range explains what governance ensemble applies. This combination produces traceable defensible decisions rather than ad hoc labels.
Framing and limits
R is a measurement tool not a policy oracle. It clarifies where governance attention should attach but policy design what to permit what to conditionally allow what to prohibit remains a jurisdictional choice. DyFrams R score is designed to make those choices informed auditable and consistent across time and deployments.
