Deepfakes are a problem on the internet because they can be used to manipulate or defame people, making them say or do things they didn’t say or do.
According to Tal Hassner and Xi Yin, two social network researchers who worked on the issue, along with Michigan State University (United States), the system “will facilitate the detection of ‘deepfakes’ and the tracking of related information”.
The method should provide “tools to better investigate incidents of coordinated disinformation that use deep fakes”, they said, quoted by French news agency AFP.
To develop the system, the researchers used a technique known as “reverse engineering”, which consists of deconstructing the manufacture of a product or, in this case, a video or photograph.
The programme identifies imperfections added during editing, which alter the fingerprint of the images.
In photography, this fingerprint can be used to identify the model of camera used; in computing, the aim is to identify the generation system used to produce the lie.
Microsoft last year introduced a programme that can help detect fake photos or videos, one of several ‘software’ designed to combat disinformation ahead of the US presidential election.
In late 2019, Google made thousands of video ‘deepfakes’ public, making them available to researchers who wanted to develop methods to detect manipulated images.