Visual aids: diagrams with text equivalents
Diagrams support facilitation. Text equivalents below remain authoritative for screen readers and PDF/UA-1 exports.
AI use decision flow
Text equivalent:
- Purpose: What task does AI support? What is the manual alternative?
- People: Who is affected? Was consultation adequate for decisions affecting them?
- Data: What data will you enter? Is the tool approved for that data?
- Tool: Where does input go? Training use? Third-party access?
- Verification: Which claims need a trusted source? How will you check?
- Human control: Who approves, overrides, and provides remedy?
- Monitoring: What is the test period? What triggers a stop?
- Decision: Record green (proceed with checks), amber (pause for review), or red (stop) with a named owner and review date.
See also: AI use decision guide.
Data responsibility lifecycle
Text equivalent (IASC programme cycle):
| Stage | Question before using AI on operational data |
|---|---|
| Plan | Is AI necessary? Who is accountable? |
| Assess | What harms could data or AI outputs cause? |
| Collect | Is collection limited, lawful, and explained to people? |
| Share | Are sharing agreements and purpose limits clear? |
| Analyse | Could analysis re-identify people or mislead decision-makers? |
| Store | Retention, access control, and breach response defined? |
| Close | When and how is data deleted or returned? |
See also: Data-responsibility checklist and IASC Operational Guidance (2023).
Safeguarding escalation (session use)
Text equivalent:
| Step | Action |
|---|---|
| 1 | Pause exercise when real sensitive information appears |
| 2 | Remove identifying content from chat per platform rules |
| 3 | Refer participant to org safeguarding, PSEA, or protection focal point |
| 4 | Do not investigate in plenary |
| 5 | Notify hosting org contact after session without copying sensitive detail |