# AI Transparency Compliance Assessment Template

## Organisation: Cerebrum AI Labs

---

## SECTION 1: AI SYSTEM IDENTIFICATION

| Field | Value |
|-------|-------|
| System Name | |
| System Description | |
| AI Act Classification | Unacceptable / High-risk / Limited / Minimal |
| Assessment Date | |
| Assessor | |
| Version | 1.0 |

---

## SECTION 2: GDPR TRANSPARENCY CHECKLIST (Arts. 13-14)

| # | Requirement | Article | Met? | Evidence / Location |
|---|------------|---------|------|---------------------|
| 1 | Specific AI processing purpose disclosed | Art. 13(1)(c) | Yes/No | |
| 2 | Lawful basis for AI training and inference | Art. 13(1)(c) | Yes/No | |
| 3 | Legitimate interest pursued via AI | Art. 13(1)(d) | Yes/No | |
| 4 | AI infrastructure recipients disclosed | Art. 13(1)(e) | Yes/No | |
| 5 | AI processing jurisdictions disclosed | Art. 13(1)(f) | Yes/No | |
| 6 | Training data/inference log retention | Art. 13(2)(a) | Yes/No | |
| 7 | AI-specific data subject rights disclosed | Art. 13(2)(b) | Yes/No | |
| 8 | Meaningful info about AI decision logic | Art. 13(2)(f) | Yes/No | |
| 9 | Training data sources disclosed | Art. 14(2)(f) | Yes/No | |

---

## SECTION 3: AI ACT TRANSPARENCY CHECKLIST

### Art. 50 — General Transparency

| # | Requirement | Applicable? | Met? | Implementation |
|---|------------|-------------|------|----------------|
| 1 | AI interaction notification (Art. 50(1)) | Yes/No | Yes/No | |
| 2 | Emotion recognition notice (Art. 50(2)) | Yes/No | Yes/No | |
| 3 | AI content labelling (Art. 50(3)) | Yes/No | Yes/No | |
| 4 | Public interest text disclosure (Art. 50(4)) | Yes/No | Yes/No | |

### Arts. 13-14 — High-Risk AI (complete only if high-risk)

| # | Requirement | Met? | Evidence |
|---|------------|------|----------|
| 1 | System designed for output interpretability | Yes/No | |
| 2 | Instructions for use provided | Yes/No | |
| 3 | Performance metrics documented | Yes/No | |
| 4 | Known limitations disclosed | Yes/No | |
| 5 | Human oversight design documented | Yes/No | |
| 6 | Art. 86 individual explanation mechanism | Yes/No | |

---

## SECTION 4: LAYERED TRANSPARENCY IMPLEMENTATION

### Layer 1: Initial AI Notification

| Element | Status | Details |
|---------|--------|---------|
| Notification text | | |
| Placement (where displayed) | | |
| Timing (when displayed) | | |
| Accessibility tested | Yes/No | |
| Languages available | | |

### Layer 2: Privacy Notice AI Section

| Element | Included? | Location in Notice |
|---------|-----------|-------------------|
| AI system description | Yes/No | |
| Processing purpose | Yes/No | |
| Lawful basis | Yes/No | |
| Data categories (training + inference) | Yes/No | |
| Decision authority level | Yes/No | |
| Key factors in AI output | Yes/No | |
| Data subject rights | Yes/No | |
| Retention periods | Yes/No | |
| Recipients | Yes/No | |
| International transfers | Yes/No | |

### Layer 3: Detailed AI Documentation

| Document | Available? | Location | Last Updated |
|----------|-----------|----------|-------------|
| Model card | Yes/No | | |
| Training data description | Yes/No | | |
| Fairness and accuracy metrics | Yes/No | | |
| Known limitations document | Yes/No | | |
| Human oversight procedures | Yes/No | | |
| Privacy safeguards description | Yes/No | | |

### Layer 4: Individual Decision Explanation

| Element | Status | Details |
|---------|--------|---------|
| Art. 22 applicable? | Yes/No | |
| Explanation mechanism deployed | Yes/No | |
| Explainability tool | SHAP / LIME / Counterfactual / Other: |
| Response time for explanation requests | | |
| Human review escalation available | Yes/No | |
| Contestation mechanism | Yes/No | |

---

## SECTION 5: MODEL CARD

| Section | Content |
|---------|---------|
| Model name and version | |
| Model type/architecture | |
| Developer | |
| Deployment date | |
| Intended use | |
| Out-of-scope use | |
| Training data (categories) | |
| Training data volume | |
| Performance metrics | |
| Fairness metrics | |
| Known limitations | |
| Privacy properties | |
| Human oversight requirement | |
| Last updated | |

---

## SECTION 6: GAP ANALYSIS AND REMEDIATION

| # | Gap Description | Framework | Article | Severity | Remediation Action | Owner | Deadline |
|---|----------------|-----------|---------|----------|-------------------|-------|----------|
| 1 | | GDPR/AI Act | | Critical/Major/Minor | | | |
| 2 | | | | | | | |

---

## SECTION 7: APPROVAL

| Role | Name | Date |
|------|------|------|
| Privacy/DPO | | |
| AI System Owner | | |
| Legal | | |

**Next Review Date**: _______________