Spotting AI-generated images and videos

AI-generated images and videos are everywhere: in feeds, ads, news and group chats. Some are clearly labelled art; others are made to deceive. Helping your child spot them is part of teaching them to check what they see online.

Visual clues (less reliable over time)

Earlier AI images often had obvious mistakes. Some still do:

• Hands and fingers that look odd
• Text in the image that's garbled or nonsensical
• Earrings, glasses or teeth that don't match
• Backgrounds that blur or warp strangely
• Skin that looks too smooth, like plastic
• In videos: unnatural blinking, lips out of sync, flickering edges

But AI tools are improving fast. Don't rely on spotting mistakes. Many fakes have none.

Better: check the source

• Who posted it? A known news outlet, or an anonymous page?
• Where did it come from? Search for the original.
• Is anyone else reporting it? A real major event will be covered by multiple reliable sources.
• Reverse image search with Google Lens: it can show where an image appeared first, and whether it's been fact-checked.

Look for labels

• Platforms like Facebook, Instagram, TikTok and YouTube label some AI content with notes such as "AI info" or "Made with AI".
• Some AI tools add invisible watermarks or "content credentials" that verification tools can read.
• Labels help, but many fakes aren't labelled.

Pause before sharing

Teach a simple habit: if it's shocking, outrageous or too perfect, pause. Fakes are designed to trigger strong emotions so people share them fast.

Practise together

• Look at a set of images and guess which are AI. Then check.
• Try making an AI image together, to see how easy it is.
• Talk about why someone would make a fake: money, politics, attention, scams.

When fakes target real people

Fake images of real people, especially sexual ones, are harmful and sometimes illegal. See deepfakes and nudify apps.