Private?
Don’t upload it.
Skip to content
DeraSafeSafety Starts with Awareness
DeraSafeSafety Starts with AwarenessPractical guide
AI Safety & Responsible Use
Use AI to support learning, creativity and decision-making — without giving up privacy, critical thinking or responsibility.
The DeraSafe Five Rules
Important?
Verify it.
Learning?
Think before you ask.
Not your work?
Don’t claim it.
Sensitive issue?
Talk to a trusted human.
For Children
AI can make mistakes. Keep personal information private. AI-made images and voices may look real. Do not use AI to embarrass or hurt others. If something worries you, talk to a trusted adult.
For Students
Use AI to learn, not to copy. Verify important answers, protect personal information and follow academic integrity rules. Completing the task is not the same as learning.
For Parents
Ask what tools your child uses, what they share, and how they verify answers. Discuss deepfakes, privacy, academic use and AI companions.
For Educators
Use AI to support planning and review every output. Student data is not a prompt. Create clear rules and assess thinking, not only the final product.
Before You Hit Send
Do not upload names, identification numbers, phone numbers, addresses, images of children, health information, school records, passwords, verification codes, financial information or another person’s data unless there is a necessary, authorised reason.
If a text or file includes private information, stop and remove or anonymise it before using an AI tool.
Is My AI Use Responsible?
Use the checklist before sending or submitting work.
Family AI Agreement
- We do not share passwords, verification codes or unnecessary private information.
- We verify important information and do not use AI to hurt or deceive others.
- We do not submit schoolwork we do not understand as our own.
- We tell a trusted adult if AI content is worrying or inappropriate.
Classroom AI Policy
Allowed: explain, practise and brainstorm when the task permits it.
Allowed with conditions: use AI in named stages and disclose it where required.
Not allowed: submit AI-generated work as personal work when a task measures individual performance.
AI Hallucinations & Verification
AI can produce answers that sound confident but are inaccurate or invented. Ask for the source, open the original source yourself and check the date and context.
Deepfakes & AI-Generated Content
Do not rely on appearance alone. Check the original source, look for credible reporting and do not reshare attention-grabbing content before verifying it.
NCMEC reported more than 400,000 CyberTipline reports with a generative-AI nexus in 2025, including more than 182,000 involving possession, generation or attempted generation of GAI CSAM. It also noted cases involving “nudify” tools used to create harmful manipulated imagery, including among peers and classmates.
View carefully contextualised data →When to Talk to a Trusted Human
AI is not a substitute for a parent, teacher, doctor, counsellor, lawyer or official authority in a sensitive situation.
When Something Goes Wrong