01
It predicts, it does not know
A chatbot works out the most likely next words. That makes it sound fluent and sure of itself, and it also means it can be completely wrong while looking completely right.
— AI for All · Staying safe —
AI safety for young people,
explained simply.
AI safety for young people does not have to be frightening. AI for All by the Institute of AI gives young people, and the adults around them, free and plain guidance on using AI tools sensibly, spotting the risks, and staying in control online.
— Start here —
The calm version,
in three ideas.
AI tools like chatbots and image generators can genuinely help you learn and create. They are not magic, and they are not out to get you. Before any rules, it helps to hold three plain facts about how they actually work.
02
It remembers what you tell it
What you type can be stored, and in some tools it is used to improve the model. Treat every box you type into as if it were public, not a private diary.
03
You stay in charge
AI is a tool, not a teacher, a friend, or a source of truth. The person deciding what to trust, what to share, and what to do next is still you.
— Rules and red flags —
Sensible habits,
and the risks to spot.
Two simple lists. Good habits worth keeping on one side, and the warning signs worth knowing on the other. Learn both and you can use almost any AI tool with confidence.
Good habits
Sensible rules to keep
Check before you trust it
AI can sound certain and still be wrong. Look anything important up somewhere you already trust before you rely on it, share it, or hand it in.
Keep your personal information to yourself
Your full name, address, school, passwords, photos, and anything about your family are yours to keep. Do not type them into a chatbot or an app, however friendly it feels.
Remember it can be confidently wrong
A fluent answer is not a checked one. Treat what a tool tells you as a first draft to verify, never as the final word.
Tell an adult if something feels off
If a tool asks for something odd, shows you something upsetting, or a conversation starts to feel wrong, stop and tell a parent, carer, or teacher. You will not be in trouble for asking.
Your work should still be your own
AI is fine for ideas and feedback, but the thinking and the words you hand in should be yours. Check your school or college rules, and when in doubt, ask.
Red flags
Risks to spot
Hallucinations
AI invents facts, quotes, links, and sources that look completely real. If you cannot find it anywhere else, treat it as made up.
Deepfakes and scams
Photos, voices, and videos can be faked, and messages can pretend to be someone you know. Slow right down before you believe, click, or pay.
Oversharing
Chatbots and apps can keep what you tell them. The safest rule is to share nothing you would not pin to a noticeboard at school.
Over-reliance
Leaning on AI for everything can quietly weaken your own skills. Use it to help you learn, not to skip the learning.
For parents and teachers
You do not need to be technical to keep them safe.
The most useful thing an adult can do is stay curious alongside a young person, not police every click. Ask what tools they use, look at them together, and agree a few plain rules for your home or classroom. AI for All by the Institute of AI publishes free, plain-English guidance you can share, and the AI glossary explains any term you meet along the way.
Agree where AI helps
Decide together where AI is welcome in schoolwork and where it is not, so the line is clear before it is tested.
Set things up together
Create accounts side by side, look at the privacy settings, and check the age guidance on each tool before it is used.
Keep the conversation open
The most useful habit is talking, so a young person comes to you first when something online feels wrong.
— Common questions —
Questions about AI safety for young people.
Is it safe for young people to use AI chatbots?+
Used sensibly, yes. Stick to tools that are meant for your age, keep personal information to yourself, check anything important before you trust it, and tell an adult if something feels wrong. The good habits on this page cover the essentials without needing any technical knowledge.
What are the main risks of AI for young people?+
Four are worth knowing well: hallucinations, where a tool invents facts that look real; deepfakes and scams, where photos, voices, and messages are faked; oversharing, where apps keep what you tell them; and over-reliance, where leaning on AI too much weakens your own skills. Spotting these is most of staying safe.
Can I use AI for my schoolwork?+
It depends on your school or college rules, so check them first. As a general guide, AI is fine for ideas, explanations, and feedback, but the thinking and the words you hand in should be your own. If you are unsure whether something is allowed, ask a teacher rather than guess.
How can I tell if something is an AI deepfake or a scam?+
Slow down and be suspicious of anything that rushes you to believe, click, or pay. Check whether the same claim appears anywhere you already trust, look closely at odd details, and be wary of messages that only feel like someone you know. When in doubt, do not act, and tell an adult.
Is this AI safety guidance free?+
Yes. AI for All is the public programme of the Institute of AI, and its guidance for young people, schools, and charities is free to read and share. You can go deeper any time through Learn AI, including the AI glossary and plain-English lessons.
Stay curious,
and stay in control.
AI for All by the Institute of AI keeps its guidance free and plain, for young people and the adults around them. Explore it, and keep learning as the tools change.
— Go deeper
— AI for All
AI for All: Young people
Free guidance, events, and resources for students and young people, from the public programme of the Institute of AI.
— Skills
AI skills for young people
Move from staying safe to getting good: practical, age-appropriate ways to build real AI skills that last.
— Learn AI
The AI glossary
Plain-English definitions of the AI terms you will meet, from hallucination and deepfake to model and prompt.