Spotting deepfakes and AI-made images
AI can make photos, videos and voices of things that never happened. Looking closely helps a little. Checking the source helps a lot more.
In 60 seconds
Spotting deepfakes and AI-made images
AI can make photos, videos and voices of things that never happened. Looking closely helps a little. Checking the source helps a lot more.
Clues you can look for
| Where to look | Possible clue |
|---|---|
| Hands, teeth, ears, jewellery | Wrong number of fingers, melted shapes, earrings that don't match |
| Text in the image | Signboards or labels with strange, broken letters |
| Lips and voice | Mouth movement doesn't match the words; voice sounds flat or oddly smooth |
| Eyes and face edges | Blurry edges around the face, odd blinking, skin that looks too smooth |
| Light and shadows | Shadows pointing in different directions |
| Background | Bent lines, repeated patterns, people with merged faces |
Better: check the source
- 1
Who posted it first?
An official account, a known news outlet, or a random forward? - 2
Is anyone trusted reporting it?
If a famous person really said something shocking, several reliable news sources will cover it. - 3
Search for the image
Many search apps let you search with an image. You may find the original, older photo. - 4
Check fact-checkers
Fact-checking organisations often check viral posts. Search the claim with the word "fact check". - 5
Wait before sharing
If you can't confirm it, don't forward it.
- "Forwarded many times", so it must be true
- Sharing because it makes you angry
- "I'll share it just in case"
- Checking who posted it first
- Pausing when a post makes you very angry or scared
- Not sharing what you can't confirm
Watch and read more
Lab
A checked verdict on one real viral post.
The problem
Find one viral image or video you received recently (or pick a popular one from the news).
- Write down what the post claims.
- Look for visual clues from the table in this lesson and note what you see.
- Check the source: who posted first? Do trusted news sources or fact-checkers mention it? Try an image search if your phone has one.
- Write your verdict: confirmed, false, or can't tell, with your reasons.
Don't forward the post while you check.
You are done when
Hard questions
Try to answer before you reveal. If you can answer these, you understood the lesson.
Q1Why does "look for mistakes" become weaker advice over time, while "check the source" stays strong?Reveal
AI image and video tools keep improving, so the visible mistakes disappear. But a fake still has to come from somewhere, and trusted sources still won't confirm something that didn't happen. Source-checking tests the story, not the pixels, so it does not depend on how good the fake looks.
Q2If people learn that anything can be faked, what new danger appears?Reveal
People may stop believing real evidence too, and anyone caught on real video can simply say "that's AI". This makes trusted sources, careful journalism and ways to prove where media came from even more important.
Questions people ask
Are there apps that can detect all deepfakes?
No tool catches all of them. Detection tools can help experts, but they make mistakes both ways. For ordinary people, checking the source is more reliable.
Is making AI images always wrong?
No. People use AI images for art, school projects, posters and fun. It becomes harmful when it is used to deceive, cheat, harass, or pretend to be a real person without permission.
My friend's photo was used in a fake. What should we do?
Don't share it. Take screenshots and save links as proof. Report it on the platform where it appeared. In India you can report at cybercrime.gov.in or call 1930. Support your friend; it is not their fault.
Can a real video be called fake?
Yes. Some people now say "it's a deepfake" about real evidence to avoid blame. This is another reason to check sources carefully rather than trusting feelings.
Lesson test
5 questions. Get 3 right (60%) to pass and complete this lesson.