AIGH TRANSACTION RISK MODEL
TRM-1 makes risk signals explainable before a human decides.
An AIGH-owned, reproducibly trained model for governance research. It accepts seven non-identifying operational signals, returns a synthetic risk estimate, and never authorizes, blocks, or declines a payment.
Synthetic validationNo card or identity dataHuman decision required
03 · GOVERNANCE BOUNDARY
This is a model demonstration—not a payment control.
Permitted: synthetic evaluation, explainability review, threshold research, red-team testing.
Prohibited: real customer data, live authorization, automated denial, production routing, compliance claims.
Required before pilot: legal basis, data minimization, bias testing, monitoring, human appeal, rollback, and payment-provider approval.