Can you believe everything you see in a video? Less and less. Deepfakes, fake videos and images created with artificial intelligence, have become so realistic that we no longer know how to tell the authentic from the invented. We tell you what they are, how they work and, above all, how to protect yourself.
What exactly a deepfake is
A deepfake is audiovisual content manipulated with artificial intelligence so that a person appears to say or do something they never said or did. The word combines deep learning and fake.
The technique is not new: photo retouching has been with us for decades. The difference is that the tools are now so powerful that anyone can create a convincing fake from their phone, without being an expert.
How they are created
To make a deepfake, a neural network analyses thousands of images and videos of a person: their gestures, their way of speaking, their expressions. With that material, it learns to imitate them.
Then, the system superimposes that “learned face” onto another body or makes it say an invented text. The result is a video where the mouth, the voice and the gestures seem real, even though they never happened.
The good uses… and the bad ones
Not everything is negative. The same technology serves to restore the voice of people who have lost the ability to speak, to dub films naturally or to create characters in video games and special effects.
But it also has a dark side: political disinformation, identity theft in scams and, above all, fake intimate videos that destroy the reputation of real people. It is one of the most harmful and hardest uses to stop.
How to detect a fake video
Although they keep getting better, deepfakes usually leave clues. Look at these details:
- Blinking: the eyes move strangely or barely blink.
- The mouth: the lips do not sync well with the sound.
- The edges of the face: there is usually a blurry outline or changes in lighting around the face.
- The skin and hair: they look too smooth or with an artificial shine.
- The hands and teeth: they remain difficult to reproduce accurately.
If something seems suspicious to you, stop the video and look for the original source. A good reflex is to ask yourself: does it make sense for this person to say this? Have reliable media confirmed it?
The tools to fight against them
Technology is also used to defend itself. There are programmes that analyse imperfections invisible to the human eye, such as micro-movements of the face or the digital fingerprints left by each camera.
In addition, many platforms already label or remove manipulated content, and some governments are working on laws to force marking videos generated by artificial intelligence.
Your best defence: critical thinking
The golden rule is simple: distrust what surprises you too much. A controversial, shocking video or one that confirms your ideas is precisely the one that can most easily be fake.
Always check the source, look for the same content in trusted media and do not share anything without verifying. In the age of deepfakes, seeing is no longer the same as believing.
