How to tell if a video was made with AI: 7 checks to avoid falling prey to fraud
2026-03-05 13:16:22
TiranaPost
Protecting our digital identity has become increasingly important, especially in a world where content created by Artificial Intelligence is spreading rapidly. Among them, one of the most widely used techniques to manipulate faces, voices and behavior is deepfake video. But what exactly is it and how can we tell if a video is fake?
The term deepfake, coined in 2017, comes from the combination of the English words “deep learning” (an artificial intelligence technique that mimics the way the human brain learns) and “fake.” The aim of this technique is to combine or overlay real images, audio, and video with artificially generated material, using advanced machine learning methods.
The result is fake videos, audio and photos that aim to manipulate reality and simulate events or people that never existed. Often the protagonists of these videos are public figures, who appear to say or do something that they never actually did. The aim may be to influence public opinion, create conflicts, damage the reputation or defame the people involved.
Deepfake technology is also used to spread fake pornographic videos, for online fraud, for online bullying, satire or even other cybercrimes. Over the years, these videos have become increasingly realistic, often making it difficult to distinguish between a real video and a manipulated one. However, the consequences can be very serious for the victims, causing psychological, social and economic problems.
However, identifying them is not impossible. Here are some simple ways to tell if a video is a deepfake:
1. Check the source of the publication
The first element to look at is the source: the page or profile that published the video. If the video does not come from a well-known media outlet, an official profile, or a reliable source, but from a suspicious page or an account without a reputation, there is a possibility that it is a deepfake.
2. Check if the lip movements match the voice.
Often in deepfake videos there is a discrepancy between the movement of the lips and the words that are heard. You may notice delays, a robotic tone, unnatural pauses, or a monotonous way of speaking.
3. Look for strange visual details
Other signs may include unusual eye movements, lack of eyelash fluttering, blurred areas on the face, strange shadows, or elements in the background that appear distorted.
4. Check metadata and AI watermarks
Some artificial intelligence technologies place tags or marks that indicate a video was created or edited with AI. The file's metadata can also indicate whether the material has been manipulated.
5. Compare it with other reliable sources
If the video shows an important event, it is very likely that other media outlets have also reported it. If there is no confirmation from reliable sources or fact-checking sites, manipulation should be suspected.
6. Notice the person's behavior and expressions.
In deepfake videos, facial expressions may not match the words, reactions may seem strange, and movements may be unnatural.
7. Use online verification tools
When in doubt, you can use fact-checking platforms or digital tools that analyze images and videos to detect manipulation.
With the spread of artificial intelligence technology, deepfake videos are becoming increasingly common. Therefore, knowing how to spot them is an important way to avoid being scammed online.
