Teaching a child to spot an AI deepfake starts by changing the question in their head from "does this look real?" to "where did this actually come from?" — because fake images and clips are now cheap to make and often too convincing for the eye alone.

What is a deepfake, and why does my child need to know

A deepfake is an image, voice or video generated by AI to make it look like a real person said or did something that never happened. It used to take a skilled team days to produce one; now anyone can do it in minutes with a free app. I have seen students paste a classmate's face into a joke clip just for fun, which sounds harmless, but it is the exact same technique scammers use to trick adults out of money. A child who understands this early will not fall for it, and will not accidentally hurt someone else with the same tool.

How can you tell a picture or clip was made by AI

The old tricks, like counting fingers or checking uneven teeth, no longer work reliably in 2026. Newer models have fixed most of those tells, so I have stopped teaching children to hunt for visual glitches. The smoother an image looks, the more confidently wrong that method gets. What still works is checking the context around the image, not the image itself.

Four checks before you believe anything

  • Who posted it — does this account have a history and a real identity, or did it appear a few days ago
  • Do trusted outlets have the story — if something genuinely big happened, real news organisations report it too, not just one lone clip
  • Is it asking for money or urgency — pressure to act fast is the biggest red flag of any scam content
  • Reverse-search it or ask an adult — train your child to pause for one beat and check with a parent or teacher before sharing or believing anything
Do not train your child to spot a fake picture. Train them to question where the picture came from.

Scammers now clone voices too — give your family a code word

A cloned voice can be more dangerous than a fake picture, because a few seconds of audio is enough to imitate a family member convincingly. Scammers in Thailand already call using a cloned relative's voice, asking for an urgent transfer. I recommend every family set up a household code word — a phrase used to confirm identity whenever someone calls asking for money or emergency help. If the caller cannot give the code, hang up and call the saved number back directly. Make sure your child knows they can ask for the code too, not only adults.

What if your child wants to make one themselves

Plenty of children want to try generating images or clips with AI, and that curiosity is healthy in the right setting. I draw one clear line with my students: making something for yourself or friends who already know it is a joke is fine, but putting someone else's face into a situation they never agreed to is not, even as a joke, because the line between funny and hurtful is thin and children rarely notice when they have crossed it. Learning to prompt AI image tools responsibly should go hand in hand with learning to prompt them well in the first place, which is exactly what the prompting ChatGPT to draw what you want course covers.

For a fuller picture of staying safe online while learning to code, read our article on online safety for young coders, or learn alongside your child in the AI literacy course for parents. If you want to build real cybersecurity awareness, the cybersecurity basics course is open from age 10, and if you are unsure where to start, our course advisor page can point you the right way.

Frequently asked questions

At what age should a child learn about deepfakes

Roughly from the time they have their own social media account or watch clips unsupervised, usually from around age 9 or 10. Before that, a parent should be filtering content for them.

What should my child do if they find a fake clip with a friend's face in it

Tell them to screenshot it as evidence, not share it further, and report it to a parent or teacher straight away. This is not something a child should handle alone.

Are the AI image apps my child plays with dangerous

The apps themselves are not the danger — how they are used is. Keep them within your sight and agree on boundaries before your child starts using one, not after a problem happens.