Zuletzt aktualisiert am:
Synthetic media in which a person in an existing image or video is replaced with someone else's likeness, created using deep learning techniques.
Generated via GANs or autoencoders that learn to map one face onto another convincingly. Deepfakes raise governance concerns - misinformation, defamation, fraud - and drive the development of detection algorithms, watermarking standards, and legal frameworks to label or ban malicious uses.
Praxisbeispiel:
A political campaign is targeted with a deepfake video showing a candidate making inflammatory remarks. The campaign’s forensic-AI team uses a detection model that spots inconsistencies in eye blinking and face textures, verifies the video as fake, and issues a public correction within hours.
