Deepfake
A deepfake is image, audio or video content made or altered with AI so that it convincingly shows a real person, place or event in a way that did not happen and could pass as authentic.
Also called: synthetic media, deep fake
By Divya Sharma · Checked Nov 15, 2026 · How we review
The term started with face-swapped videos, but it now covers any realistic fake: a cloned voice on a phone call, a video of a politician saying words they never said, or an invented photo of a real event. The EU AI Act defines a deep fake as AI-generated or manipulated content that resembles real persons, objects, places or events and would falsely appear authentic.
How they are made
Most deepfakes combine generative tools that are also used for legitimate work: voice cloning for speech, lip-sync models to match mouth movement to new audio, and text-to-video or face-swap models for footage. The same tools power dubbing, accessibility and film effects, so the harm lies in deception, not in the technique.
Deepfake versus synthetic media
Synthetic media is the wider term for any AI-generated image, audio or video, including openly labelled art and stock imagery. A deepfake is the subset made to pass as real and usually tied to a real person or event.
Responses
Two approaches work alongside each other. Labelling marks content as AI-made at the source, through AI watermarking or metadata. Provenance, such as C2PA content credentials, lets honest publishers show where a file came from and how it was edited, rather than trying to prove that a file is fake. The EU AI Act requires deepfakes to be disclosed as artificial, with some exceptions. Neither labels nor detectors settle the question alone, so check the source of a surprising clip before sharing it.
Related terms
- C2PA content credentials
- C2PA content credentials are signed provenance records, defined by an open standard from the Coalition for Content Provenance and Authenticity, that travel with an image, video, audio file or document and record its origin and edits.
- AI watermarking
- AI watermarking is embedding a hidden, machine-detectable signal in text, images, audio or video at the moment an AI model generates it, so software can later identify the content as AI-generated.
- Voice cloning
- Voice cloning is the use of AI to make a text-to-speech system speak in a specific person's voice, learned from a recording of that person. Some systems need only a short sample.
- EU AI Act
- The EU AI Act, Regulation (EU) 2024/1689, is the European Union's law on artificial intelligence. It sets risk-based rules for those who build and deploy AI systems, from outright bans to transparency duties, phased in from 2025.
- Lip sync (AI video)
- AI lip sync is the use of a model to change the mouth movements of a person in a video or image so they match a new audio track, such as a translated or rewritten line.
- Text-to-video
- Text-to-video is the task of generating a short video clip from a written prompt. Models must produce frames that each look right and also move consistently from one frame to the next.