Session 1 learning objectives
Level 1: Responsible AI foundations | 90 minutes
By the end of this session, participants will be able to:
- Distinguish generative AI from prediction and automation tools, and state what each can and cannot do in humanitarian work.
- Explain who remains accountable when AI supports a humanitarian decision.
- Compare an AI-assisted draft to an authorized source and mark claims as supported, contradicted, or not established.
- Apply a seven-step risk screen to a fictional case and record a green, amber, or red decision with a named owner.
- Identify at least one protection or localization issue in a synthetic communications scenario.
- Choose one safe action to apply before the next session.
Observable behaviours (assessment alignment)
| Objective | Observable behaviour |
|---|---|
| Distinguish AI types | Correctly classifies an example tool as generative, predictive, or automation in a recall question |
| Accountability | Names a human decision owner, not the tool vendor |
| Source comparison | Completes a source-check table with at least one contradiction flagged |
| Risk screen | Records G/A/R with a reason tied to data, verification, or control |
| Protection | Describes who could be harmed by a wrong message or omission |
| Action plan | Writes one specific, low-risk next step |