AI Club Workshop Facilitator and Learning-Evidence Pack

Prepare a facilitator-ready AI club workshop that connects a specific learning objective to hands-on activity and proportionate evidence of learning.

## Workshop inputs

Audience, prior knowledge and access needs:
{{audience_and_access_needs}}

Topic, learning objectives and session constraints:
{{topic_outcomes_and_constraints}}

Approved sources, examples and tools:
{{sources_examples_and_tools}}

Institutional rules, data boundaries and available support:
{{rules_and_support}}

## Working rules

- Separate source-backed explanation, facilitator interpretation, assumption, conflicting evidence, missing information and unresolved uncertainty.
- Cite supplied sources for consequential technical or research claims. Never invent a citation, tool capability, workshop result or participant response.
- Use synthetic, open or explicitly authorized material in exercises. Do not request passwords, confidential work, identifiable learner records or restricted institutional data.
- Match activities to the named audience and available devices, bandwidth, time and accessibility needs. Provide a low-bandwidth alternative where a live tool is optional.
- Use pre/post evidence to assess the stated learning outcome, not to grade participants or infer broad ability from one activity.
- State any facilitator decision, safety pause or escalation role explicitly. Do not imply institutional endorsement or approval.
- Record the model and version, account or access tier, enabled tools and material feature differences for every exercise. Set explicit cost limits and provide a no-cost or offline alternative where practical.
- Plan for refusals, rate limits, unavailable tools and other service failures without asking learners to weaken safeguards. Use synthetic exercise data by default.

## Build the session

1. Convert the topic into two to four observable learning outcomes and identify misconceptions or risky shortcuts the session must surface.
2. Select the minimum source set. Create a claim-source card for each essential explanation, noting publication date, applicability and dispute.
3. Design an opening probe that captures prior reasoning without collecting sensitive data. Make it comparable with the closing task.
4. Sequence short explanations, demonstrations and participant exercises. For each activity specify purpose, time, model/version, account tier, enabled tools, cost ceiling, synthetic data, instructions, expected evidence and facilitator observation.
5. Add challenge cases involving missing evidence, conflicting outputs, privacy boundaries or unsupported claims where relevant to the topic.
6. Prepare facilitation cautions: likely misconception, accessibility adaptation, refusal handling, rate-limit or tool-failure fallback, cost stop, discussion boundary and when to stop an unsafe exercise.
7. Design the closing transfer task and a lightweight comparison method. Distinguish participation, correct recall, reasoned application and unresolved learning need.
8. Provide follow-up resources and a retention or application check that can be run later without requiring an account.

## Output contract: Workshop and Learning-Evidence Pack

Return:

1. **Session card**: audience, prerequisites, outcomes, length, tools, access assumptions and data boundaries.
2. **Source-grounded facilitator notes**: key claim, source reference, explanation, limitation and likely misconception.
3. **Run of show**: timed segment, facilitator action, participant action, materials, evidence produced and fallback.
4. **Exercise sheets**: instructions, synthetic or authorized inputs, expected reasoning, extension and accessibility alternative.
5. **Facilitator caution register**: risk, trigger, response and responsible role.
6. **Before-and-after evidence plan**: matched probe, observable indicators, interpretation boundary and anonymous aggregation method.
7. **Learning review**: outcome, evidence expected, success condition, unresolved need and follow-up action.
8. **Resource list**: dated source, why it matters, access status and suggested next practice.

## Completion conditions

The pack is complete only when each outcome has a corresponding activity and observable evidence; sources and permissions are traceable; examples are safe to share; access alternatives are specified; facilitator cautions cover foreseeable risks; and the closing task tests transfer rather than confidence alone.

For each completion condition, record the expected learning or permission evidence, the actual supplied evidence, `Met`, `Not met`, or `Blocked` status, and the unresolved facilitator action.

If learning objectives, sources, audience information or tool access are missing, return a scoped workshop skeleton and evidence request. Refuse to invent participant results, certify competence or claim institutional approval.