Deepfake

A deepfake is synthetic media, usually video or audio, that convincingly puts words or actions onto a real person. AI studies existing footage or recordings of someone, learns how they look and sound, and then generates new material in which they appear to do something that never happened.

The term combines “deep learning,” a form of AI, with “fake.” Early deepfakes were crude and easy to spot. As the technology has improved, some have become hard to distinguish from genuine recordings, especially in short clips or low-quality audio where small flaws are easier to hide.

To make this concrete, imagine getting a phone call in the panicked voice of your son or daughter, saying they have had an accident and need money sent right away. A scammer can clone that voice from a few seconds of audio lifted from social media, then use it to pressure you into paying before you think to check. The same trick appears as fake videos of celebrities endorsing an investment or a giveaway that they have nothing to do with.

This matters to you as a reader and viewer more than as a maker. Deepfakes are used in scams, misinformation, and impersonation, so a healthy scepticism is now part of everyday media literacy. If a video of a public figure seems designed to provoke a strong reaction, it is worth pausing to check whether reputable sources are reporting the same thing.

A few habits help. Look for the source rather than trusting a clip on its own, be wary of unexpected voice or video messages that ask for money or urgent action, and remember that “I saw it with my own eyes” no longer settles the matter the way it once did. Awareness is the strongest defence.