Artificial intelligence has flooded the ‘for you’ of all platforms and social networks. Its enormous development and improvement have meant that more and more let’s question everything we seeand sometimes, not even the most experts are able to differentiate with certainty if what they are seeing is a content generated with artificial intelligence or not.
The British media BBC has interviewed a computer science professor at the University of California, Berkeley, to ask him about the traits that help any Internet user recognize that a video or image is unreal. The main signal, he indicates, is related to the quality of the video image. If the video has poor image quality, it is most likely one generated by AI.
“It is one of the first things we analyze,” he commented in the interview carried out by the aforementioned media. The professional, in addition to being a professor, is the founder of a company dedicated to the detection of false information called Get Real Security. This feature, which at first seems obvious, is not when the video appears to have been recorded by a home security camera or an old camera. In that case, the user does not question the quality of the video, because it is common to this type of cameras, and the probability of falling into the trap increases.
This is not the only useful trick. Matthew Stammprofessor and director of the Multimedia and Information Security Laboratory at Drexel University, disagrees with his colleague’s theory, since some AI tools They are capable of creating videos with impeccable qualityand not all videos with poor quality are fake. In his opinion, the way to identify the content lies in the observation of much more specific aspects.
Other experts interviewed assure that attention must be focused on the small inconsistencies. “It’s not about six fingers or illegible text. It’s something more subtle,” they say. Some features such as skin texture, hair movement or strange changes in the background of the image help identify what is real from what is created with artificial intelligence.
Fashion AI-generated videos
Babies and dogs
They never fail. They are the winning combination if you want to go viral. This is the reason why more and more users choose to create AI-generated content in which a dog saves a baby from a disastrous fall, protects the baby from something external, or sleeps cuddling. The emotional factor makes the user not wonder if it is actually a real video or not.
security camera
Videos of a supposed child talking to the security camera of a houseof a giant animal ringing the doorbell or things like that that, despite being unreal, make everyone who sees it doubt.
Controversial statements from celebrities
Although sometimes they are real, there are many videos circulating with statements by famous people recommending certain investment strategies, or making serious accusations about other public figures. To verify this, it is necessary to compare the information with official media and see if they have echoed the words spoken by the person, since in many cases they are false.