Real or AI? How kids can spot AI images and deepfakes

AI can now make pictures, voices and videos that look very real. Children do not need to become experts, but they do need a few habits. Here is a checklist you can turn into a game.

The detective checklist

Look carefully together and ask:

1. Hands and fingers. Are there too many, too few or odd shapes?
2. Text. Is the writing on signs, shirts or books garbled?
3. Backgrounds. Do faces or objects in the distance melt or repeat?
4. Too perfect. Is the skin too smooth, or the light too dramatic?
5. Shadows and reflections. Do they match the scene?
6. Odd details. Mismatched earrings, strange teeth, glasses that merge into the face.

These clues work less well each year as AI improves, so treat them as clues, not proof.

The rule that matters more: where did it come from?

Better than looking closely is asking questions about the source:

• Who shared it, and can we trust them?
• Do real news sources show the same thing?
• Does it want me to feel shocked, angry or in a hurry?

If it looks shocking, check before you share. That one rule protects kids and the people they talk to.

What about deepfakes?

A deepfake is a video or audio clip where AI changes a real person's face or voice. For children, the message is simple: a video of someone is not always proof, and anything that asks for money, secrets or urgent action should be checked with a grown-up.

Make it a game

Collect a mix of real and AI-made pictures (many sites label them) and have your child sort them into two piles and explain why. Then reveal the answers and talk about what fooled you. The Real or AI? lesson in Bip and the day-three activity in our free mini course are built the same way.

When something is wrong

If your child sees an image that upsets them or is used to bully someone, let them know they will never be in trouble for telling you. Save it, report it on the platform and talk about it calmly.