# AI Impact Assessment

**System name:** [Name]
**System ID:** [From AI inventory]
**Assessment author:** [Name and role]
**Date:** [Date]

## 1. System Overview

| Field | Response |
|-------|----------|
| **System name and ID** | |
| **Purpose and intended use** | |
| **AI technique and architecture** | |
| **Development stage** | [ ] Design [ ] Development [ ] Pilot [ ] Production |

## 2. Stakeholder Analysis

| Stakeholder Group | Who | How Affected |
|-------------------|-----|-------------|
| **Primary users** | | |
| **Affected individuals** | | |
| **Vulnerable groups** | | |
| **Third parties / society** | | |

## 3. Benefits and Harms

| Category | Potential Benefits | Potential Harms | Likelihood | Severity |
|----------|-------------------|-----------------|------------|----------|
| Individuals | | | | |
| Groups / communities | | | | |
| Organization | | | | |
| Society / environment | | | | |

## 4. Fairness and Bias

| Question | Response |
|----------|----------|
| What protected characteristics are relevant? | |
| Has training data been assessed for bias? | |
| What fairness metrics will be used? | |
| How will fairness be monitored post-deployment? | |

## 5. Transparency and Explainability

| Question | Response |
|----------|----------|
| Are affected individuals informed AI is being used? | |
| Can the system's outputs be explained? | |
| Is there a recourse process for affected individuals? | |

## 6. Privacy and Data Protection

| Question | Response |
|----------|----------|
| Has a DPIA been completed? | [ ] Yes [ ] No [ ] Not required |
| What personal data is processed? | |
| Legal basis for processing? | |
| How is data minimization achieved? | |

## 7. Security

| Question | Response |
|----------|----------|
| Has a security review been completed? | [ ] Yes [ ] No [ ] In progress |
| What adversarial threats have been considered? | |

## 8. Human Oversight

| Question | Response |
|----------|----------|
| What level of oversight is in place? | [ ] Human-in-the-loop [ ] Human-on-the-loop [ ] Human-over-the-loop [ ] Autonomous |
| Is oversight proportionate to risk? | |

## 9. Mitigations

| Identified Harm | Mitigation Measure | Responsible | Target Date | Monitoring Approach |
|----------------|-------------------|-------------|-------------|-------------------|
| | | | | |
| | | | | |
| | | | | |

## 10. Recommendation

**Assessment author's recommendation:**

[ ] Approve
[ ] Approve with conditions
[ ] Do not approve
[ ] Need more information

**Conditions (if any):**

**Rationale:**