# AI system model card

PrivacyBuilt Editorial | Educational working model | 3 November 2026

Use this system-level record alongside provider model documentation, not as a substitute for an impact assessment or applicable legal technical documentation.

## System identity and purpose
- System / version: [complete]
- Accountable owner and provider: [complete]
- Intended use and users: [complete]
- Prohibited or unsupported uses: [complete]
- Deployment context and affected people: [complete]
## Data and permissions
- Input, retrieval, training and output data: [complete separately]
- Provenance, licences and lawful processing: [validate]
- Sensitive data exclusions and approved channels: [complete]
- Retention, deletion, provider reuse and locations: [verify]
## Evaluation and safeguards
- Evaluation datasets and provenance: [complete]
- Accuracy, subgroup performance and limitations: [attach evidence]
- Privacy leakage, prompt injection and misuse testing: [attach evidence]
- Human review authority and fallback: [complete]
- Contestability and incident escalation: [complete]
## Decision and monitoring
- Impact-assessment references and applicable legal classification: [validate]
- Approval conditions and unresolved risks: [complete]
- Change triggers, monitoring owner and review date: [complete]