How deepfakes are made
Imagine you have a robot painter who learns how your face moves when you talk. It watches lots of videos of you, and then it draws new videos where you say things you never said before. That's kind of what happens with deepfakes, the computer uses artificial intelligence, like a super-smart robot, to copy faces and voices.
Can we always spot them?
Sometimes you can tell a deepfake is fake because it looks a little off, like when someone tries to draw a smile but forgets to add eyes. But other times, they look so real that even grown-ups get tricked! It’s like when your friend wears a disguise and pretends to be someone else, sometimes it's easy to fool you, and sometimes it takes a while to figure out the trick.
Deepfakes are getting better every day, just like how you get faster at drawing with practice. So we need to learn new tricks too! Deepfakes use computer tricks to make fake videos look real, just like when you draw a picture that looks like your friend.
Examples
- Imagine watching a video of your teacher giving a speech, but it's actually your friend's face doing the talking, that's a deepfake!
- You can use a deepfake to make it look like a famous singer is singing a song they never recorded.
Ask a question
See also
- How are deepfakes created, and what are their implications?
- How do deepfake videos trick us into believing false images?
- Why are deepfakes becoming so realistic and dangerous?
- How the pros make deepfakes I Deepfakes explained?
- Are AI deepfakes of voices as convincing as video deepfakes?