Skip to content

EU AI Act Guide (Toolkit Mapping)

This page summarizes how the toolkit maps key obligations in Regulation (EU) 2024/1689. It is a technical mapping, not legal advice.

Risk Tiers in the Toolkit

  • unacceptable: prohibited practices (Art. 5)
  • high_risk: Annex III style domains and strong rights-impact signals (Art. 6, plus controls)
  • limited: transparency-heavy obligations (Art. 50)
  • minimal: baseline/low-risk context

Articles Actively Evaluated in Compliance Checker

  • Art. 5
  • Art. 10
  • Art. 11
  • Art. 13
  • Art. 14
  • Art. 15
  • Art. 43
  • Art. 50

Checker statuses:

  • compliant
  • non_compliant
  • partial
  • not_assessed

Policy baseline:

  • Missing evidence should bias to not_assessed
  • Explicit control gaps should return non_compliant
  • Mixed/partial controls should return partial

Transparency and GPAI Coverage

Transparency module provides finding-level signals for:

  • Art. 50 disclosure indicators
  • Deepfake/synthetic media indicators
  • GPAI-related obligation triggers

GPAI module assesses Art. 51-55 signals from a dedicated GPAI model descriptor and flags systemic-risk indicators using deterministic thresholds.

Deadlines and Action Prioritization

Checklist defaults:

  • Art. 5: 0 months
  • Art. 50: 6 months
  • Art. 10/11/13/14/15/43: 24 months

Only actionable statuses are emitted as checklist items (non_compliant, partial, not_assessed).

Use this toolkit for engineering controls, evidence collection, and audit trail preparation. Final compliance determinations should be validated by legal teams.