Artificial intelligence could identify cases of deep fake on the internet

Spread the love

There is a 96% accuracy rate and it can be used in real time

Artificial intelligence could identify deep fake cases on the internet

The FaceCatcher tool has an accuracy rate of 96% and can be used in real time. (photo: Unocero)

The deep fakes are ways of using artificial intelligence to carry out video editions in which the identity can be impersonated. of a person creating a representation of their face on a different body, to create an illusion of realism.

This use of technology b>It can be seen in movies, series and advertising. However, it has also been misused by people who falsify videos of national government authorities to make the world believe that statements or attitudes have been made that did not actually occur.

The case of an alleged video of United States President Joe Biden singing during a speech is a use case that could be considered funny for < b>audience. However, this technology can be used to defraud people, as occurred with cases of pornographic videos of Hollywood actresses.

In this regard, a tool became known of software that also usesartificial intelligence to detect deep fakes to expose these malicious practices to the world.

FakeCatcher, as this technological resource is known, was introduced by Intel as a detectordeep fake in real time, so it could be used instantly to recognize which video or content is being manipulated by artificial intelligence.

“This technology can detect fake videos with accuracy 96% (…) It is the world's first real-time deep fake detector that returns results in milliseconds” , indicates the company on its official website .

How FakeCatcher works

The company states in the official launch statement that the detection method is based on finding real clues using the blood flow on people's faces as a reference.

This feature is subtle in video pixels and comes with very slight changes in skin tone. After processing the images by a system of algorithms, if these signals are detected the video will be identified as a real one, while the opposite case will be instantly listed as a fake.

As part of your presentation, it is indicated that FakeCatcher can generate an increase in the trust of the media, while in the case of social networks it could be a useful tool for identifying videos that are false or that are harmful to communities.

At the global level, this function could be used to prevent the dissemination of content that is manipulated or whose authenticity is doubted. Non-profit organizations would also participate in the application of this platform by democratizing the detection of these cases for the public.

For For their part, companies like Google that are dedicated to the application of technology, have vetoed this practice of video editing due to the misuse that has been given to it on the internet and the bad image that its application currently has at the user level.

The European Union, for its part, has also become involved in the fight against misinformation as a result of this type of content and has insisted that social media companies such as Meta and Twitter have clear measures to combat deep fakes on their platforms.

Continue reading:

Posted in Uncategorized