Overview: This utility streamlines preliminary risk assessments for AI systems by evaluating telemetry inputs against the EU AI Act risk tiers and mapping them to NIST AI RMF 1.0 controls.
Required Input Parameters:
- System Name & Department: Internal product identifiers and owning operational groups.
- Deployment Domain: Determines the statutory EU AI Act risk tier (e.g., employment triggers High Risk, social scoring triggers Unacceptable Risk).
- Model Type: Technical architecture classification (Rule-based, Predictive, or GenAI).
- Data Sensitivity Level: Evaluates privacy boundaries (Public, Confidential, or Restricted PII/PHI).
- Human-in-the-Loop (HITL): Indicates whether active human review checkpoints govern execution.
Export Options:
- PDF Document Icon: Generates and downloads a clean executive report for legal or compliance review.
- JSON Code Icon (</>): Exports raw structured data for database ingestion or GRC platform tracking.