AI EncyclopediaAISafety & rules

Deepfake

딥페이크

Fake video, photos or voices made by AI that look and sound real.

Overview

The word joins deep learning and fake. An AI is trained on a person's face and voice, then produces video or audio of that person saying and doing things they never did. It used to take professional equipment, but now a few photos and a short recording can be enough.

There are good uses, such as movie special effects and dubbing. But there is a lot of harmful abuse: scam calls that imitate a family member's voice, fake speech videos before an election, and sexual fakes made without consent. It is getting harder to tell by eye, so you need the habit of checking the source and double-checking an urgent request for money through another channel.

An everyday example

It is as if someone wearing a mask of another person's face, and copying their voice exactly, has appeared on the screen. Looking at the surface, it is hard to tell if it is really them.

Why it matters

Anyone can become a victim, so many countries are strengthening laws that punish making or spreading them. The belief that "there is a video, so it is true" no longer holds.

These concept articles were written by AI and checked again for accuracy before publishing. The English edition is translated from the Korean articles. They leave out things that change quickly, such as prices and rankings. Tell us if something is wrong. Report an errorLast updated

You