What Is a Deepfake?
A deepfake is AI-generated synthetic media — video, audio, or images — that realistically depicts people saying or doing things they never actually said or did. Deepfakes use deep learning techniques (typically generative adversarial networks or diffusion models) to create convincing fabrications that are increasingly difficult to distinguish from authentic content. The term covers face swaps in video, cloned voices, AI-generated photographs of fictional people, and manipulated audio of real people appearing to say things they never said.
Deepfake — AI-generated or AI-manipulated synthetic content — audio, image, or video — that depicts real people doing or saying things they did not do or say.
Deepfake governance has accelerated sharply. The EU AI Act Digital Omnibus added an explicit prohibition on non-consensual intimate deepfake imagery (effective 2 December 2026). The UK Online Safety Act, Australian Online Safety Amendment, and US TAKE IT DOWN Act all create criminal or civil liability for distributing non-consensual deepfake intimate content. EU AI Act Article 50 requires AI-generated synthetic content to be labelled (also from 2 December 2026 under the Omnibus amendments).
Source: EU AI Act, Article 50 and Digital Omnibus; UK Online Safety Act
Why it matters for governance
The EU AI Act introduces specific deepfake governance obligations from 2 August 2026 under Article 50: AI-generated or manipulated content that constitutes a deepfake must be labelled as artificially generated or manipulated, in a clear and distinguishable manner. The EU Digital Omnibus (May 2026) added a new prohibition on AI systems used to create non-consensual sexually explicit deepfakes. Multiple jurisdictions have enacted or are considering deepfake-specific legislation covering electoral manipulation, non-consensual intimate imagery, and fraud. Organisations must implement detection tools, labelling mechanisms, and response procedures for deepfake-related incidents.