Are AI deepfakes of voices as convincing as video deepfakes?

A deepfake is like when someone uses a computer to make it look like a person is saying something they never actually said, just like how a toy robot can talk even though it's not real.

Imagine your friend has a special voice recorder that copies the way they speak. A voice deepfake is like when someone uses that recording to make it sound like your friend is talking, even if they're not there. It’s kind of like when you use a phone app to change your voice into a robot's, cool, but you know it’s not real.

Now imagine your friend is on a video call, and someone uses special software to make it look like they’re in a different place or even doing something silly. A video deepfake is like that, it makes the person look like they're somewhere else or acting in a funny way, even if they're just sitting there talking.

Video deepfakes usually feel more convincing because you can see everything, their face, their eyes, how they move. But voice deepfakes are getting really good too! They’re like the difference between hearing your friend’s voice on the phone and seeing them in person, both fun, but one feels more real to some people.