Deepfake
Also called Deep fake, Synthetic media
Realistic AI-generated or AI-altered video, audio or images that show a real person saying or doing something they never did.
What is a deepfake?
A deepfake is synthetic media that convincingly imitates a real person: a video of a politician giving a speech they never gave, a cloned voice of a manager asking for an urgent payment, or a fabricated photo. The name combines "deep learning" and "fake".
How does it work?
Deepfakes are made with generative AI models trained on images or recordings of the target. Common techniques include swapping one face onto another in a video, syncing lip movements to new audio, generating a whole scene with a video model, and cloning a voice with text-to-speech from a short sample. Tools that once needed expertise now work from a phone.
What is it commonly confused with?
- All AI-generated media. An AI avatar made with someone's consent, or a clearly labelled synthetic presenter, isn't usually called a deepfake. The term normally means content that imitates a real person to deceive, or without their permission.
- Edited media. Ordinary editing, such as cutting a clip out of context, can mislead too, but it doesn't generate new footage of the person.
Why does it matter?
Deepfakes are used for fraud, harassment, non-consensual imagery and political disinformation, and many countries now have laws against some of these uses. To protect yourself: verify unusual requests for money or data through a different channel you already trust, be sceptical of shocking clips that only appear on one account, and look for content credentials where available. If you make synthetic media, get consent from anyone it depicts and label it clearly.
