Good AI use begins with human judgement.
GUARD gives students five checks before they trust, use or share an AI-generated result — a practical classroom method, not an externally certified standard.
GGet clear on the goal
What am I actually trying to do?
UUnderstand its limits
What could this system get wrong?
AAssess accuracy & evidence
What evidence supports the result?
RRespect privacy & people
Who could be affected?
DDecide responsibly & disclose
What should I do, disclose or escalate?
Before trusting, using, submitting or sharing anything AI-generated
Judgement is the ability to decide what to trust, what to question, what to protect, and when a human should intervene. GUARD turns it into five checks students can actually practise.
Why am I asking AI this, and what skill am I not building if it does the work for me?
What could this system misunderstand, omit, invent or get wrong?
Can I verify the important claims using reliable, independent evidence?
Could this expose, misrepresent, exclude, embarrass or harm someone?
Who needs to know AI was used, and when must I stop or ask a responsible adult?
Worked scenario: a history assignment with a confident, unsourced AI quotation — G and U pass, A fails, so the student moves to R and D rather than using the quote.
Diagnose, then practise on real decisions.
Browser-based, no accounts, no student data leaving the device. Formative practice, not an accredited assessment.
Ten quick questions surface where a student's judgement is already strong and where the five checks need practice.
Take the Diagnostic →Short, single-concept labs — bias, deepfakes, privacy, verification — each ending in one real decision.
Open Focused Labs →Full missions that walk a decision through all five checks, in order, to a disclosed outcome.
Launch Scenario Lab →Eight modules, in three levels.
- What AI Actually Is
- How Generative AI Produces Answers
- Hallucinations, Truth and Verification
- Bias, Fairness and Inclusion
- Privacy, Digital Footprints and Student Safety
- Deepfakes, Peer Harm and Incident Response
- Learning With AI Without Losing the Learning
- Human Judgement, Companion AI and Accountability
Capstone: an individual or group project explained through GUARD — a classroom AI charter, a school safe-use policy, a deepfake-awareness campaign, an evaluation of an AI learning application, or a human-intervention/escalation plan.
What a completed cohort leaves behind.
At least 80% completion, the minimum assessment score, an AI-use log and a presented capstone — issued by WhizzStep from programme evidence, not an external accreditation.
Assessment Rubric →Educators and school leaders start here.
Educators
Facilitator guide, session plans, workshop deck, speaker notes and the assessment rubric.
Open Educator Hub →Schools
Pilot proposal, safeguarding protocol, feedback instruments and an outcomes report.
Open School Pilot Kit →Delivery & cohort
Classes 8–12 (adaptable for 6–7). Four 90-minute extended sessions or eight online sessions. Up to ~40 students per school cohort. In-school, live online or hybrid.
Safeguarding
The introductory workshop requires no public AI accounts. The school retains its normal supervision, consent and safeguarding responsibilities throughout.