Presentation attack
Also called presentation attack detection, PAD
Fooling a face, voice or fingerprint check by showing its sensor a fake: a photo, a mask, or a video playing on a screen.
A biometric check has a sensor: a camera for a face, a microphone for a voice, a reader for a fingerprint. A presentation attack puts something false in front of that sensor. A printed photo, a mask, a screen playing a video and a recorded voice are all presentations.
Defences against it are usually called liveness detection. They look for signs that a real, present person is there, by asking for a movement or by analysing the picture for signs of a screen or a mask.
The term matters because it marks a boundary. These defences assume the fake has to pass through the sensor. AI-made video that is fed straight into an app or a call skips the sensor, so a check can be excellent at stopping presentation attacks and still be open to that.
For example
Someone holds a phone showing another person's video up to the camera during a selfie check.
Put it to use
Where you'll meet presentation attack in ChatGPT, Gemini and Claude, how to use it there, and the mix-up most people make.
Daily Bytes Plus is in the Android app.