Deepfake in plain English.

A deepfake is video, audio, or imagery that shows a real person saying or doing something they never did, produced or altered with AI. Early versions needed hours of footage and specialist software; by 2026 a convincing voice clone needs seconds of audio and a face swap runs in real time on a laptop. The same techniques power legitimate products — AI avatars from HeyGen and Synthesia, dubbing from Camb.AI, licensed voice clones from ElevenLabs — which is why the industry distinguishes consented synthetic media from deepfakes by whether the person agreed. The harms are fraud (fake executives approving payments), non-consensual intimate imagery, and political disinformation. Defences include watermarking and content credentials, detection tools, platform policies, and laws such as the EU AI Act's labelling requirements, none of which is sufficient alone.