Contents
- GUARD Pocket Card
- Source Verification Worksheet
- Privacy-Sharing Checklist
- Deepfake Incident-Response Card
- AI Disclosure Templates
- GUARD Decision Record
- AI-Use Log
1. GUARD Pocket Card
GGet clear on the goal
What am I trying to learn, achieve or protect?
UUnderstand its limits
What could the AI misunderstand, omit, invent or get wrong?
AAssess accuracy and evidence
Can important claims be independently verified?
RRespect privacy and people
Who could be exposed, excluded, misrepresented or harmed?
DDecide responsibly and disclose
Use, revise, verify, reject, disclose, escalate or ask an adult?
Responsible action options
Use when the output is appropriate and checked. Revise when human judgement can correct it. Verify further when evidence is insufficient. Reject when it is unreliable or inappropriate. Disclose the actual assistance. Escalate when safety, rights or accountability require a responsible adult.
Stop and ask
- Could someone be harmed?
- Am I sharing someone else’s information?
- Should a responsible adult intervene?
2. Source Verification Worksheet
| Check | Notes |
|---|
| What source did the AI or post claim? | |
| Does the source exist? | |
| Does it actually support the claim? | |
| Is there reliable independent corroboration? | |
| Could the sources be copying one another? | |
| What remains uncertain? | |
Final classification
□ Verified
□ Partly supported
□ Unsupported
□ Fabricated
□ Not yet verifiable
Confidence and reason:
3. Privacy-Sharing Checklist
| Information or action | Generally safe | Permission required | High risk / unsafe | Adult or school approval |
|---|
| Your own non-identifying idea | □ | □ | □ | □ |
| A class photograph | □ | □ | □ | □ |
| A voice recording | □ | □ | □ | □ |
| Live location or travel details | □ | □ | □ | □ |
| School records or marks | □ | □ | □ | □ |
| Health or counselling information | □ | □ | □ | □ |
| A classmate’s work or notes | □ | □ | □ | □ |
| Passwords or account access | □ | □ | □ | □ |
| Content from a shared device | □ | □ | □ | □ |
| Names, badges or school identifiers | □ | □ | □ | □ |
Before sharing
- Is the information mine to share?
- Do the people involved understand where it may go and how long it may be stored?
- Can I complete the task using less information or a non-identifying example?
- Do school rules or a responsible adult need to approve it?
4. Deepfake Incident-Response Card
1
STOP
2
DO NOT RECIRCULATE
3
PRESERVE PRIVATELY
4
VERIFY
5
REPORT
6
SUPPORT
Do
- Keep the original file, link, timestamp and context privately if safe.
- Report through the school’s designated safeguarding or incident channel.
- Ask whether the full source, provenance or context is available.
- Support people targeted by the content.
- Follow responsible adult or emergency instructions.
Do not
- Forward it “just to ask if it is real”.
- Run one detector and treat the score as a verdict.
- Confront the alleged creator publicly.
- Delete all evidence before an authorised review.
- Turn an uncertain incident into entertainment.
Immediate danger or serious harm: stop the exercise and contact a responsible adult or the school’s safeguarding lead. This card is not an emergency service.5. AI Disclosure Templates
Short disclosure
“I used an AI tool for initial ideas. I checked the information and created the final response myself.”
Medium disclosure
“I used an AI tool to generate questions and suggest an outline. I verified important claims using independent sources, rejected unsupported content, and wrote and edited the final submission myself.”
Detailed disclosure
“I used [tool or feature] for [purpose]. It contributed [specific assistance]. I checked [claims, sources, calculations, code or quotations] by [verification method]. I changed or created [student contribution]. No personal or third-party information was entered / permission and school rules were checked. I remain responsible for the final work.”
Adjust the statement
A disclosure must describe what actually happened. It does not make prohibited substitution acceptable and does not remove the duty to verify the final work.
6. GUARD Decision Record
Date: ____________________ Subject / activity: __________________________________
Task or decision:
G — Get clear on the goal
What is the real goal?
U — Understand its limits
What could the AI misunderstand, omit, invent or get wrong?
A — Assess accuracy and evidence
What evidence did you examine, and what remains uncertain?
R — Respect privacy and people
Who could be affected, and what privacy, fairness, safety or well-being concerns exist?
D — Decide responsibly and disclose
□ Use □ Revise □ Verify further □ Reject □ Disclose □ Escalate / ask an adult
Reasoning:
What should be disclosed, and to whom?
Who remains responsible?
What new evidence would change your decision?
7. AI-Use Log
| Date | Subject / assignment | AI tool and purpose | What AI contributed | What I verified and changed | Disclosure / responsibility |
|---|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
Privacy reminder: Do not record names, passwords, health details or confidential school information.