To combat this practice, it is necessary to go

AEO Service Forum Drives Future of Data Innovation
Post Reply
soniya55531
Posts: 93
Joined: Sun Dec 15, 2024 5:13 am

To combat this practice, it is necessary to go

Post by soniya55531 »

Far beyond the creation of stricter legislation; all members of society must be committed to mitigating it, whether by warning about the new modality, by disseminating correct information or by developing software that can identify and block the dissemination of this content.


Last year, Buzzfeed published a video in which former President Barack Obama talks about things he would never say, including insulting Donald Trump. As the video progresses, the user realizes that it is a deepfake. According to the website itself, the intention of the video was to warn about the dangers of this practice. Facebook – after its creator, Mark Zuckerberg, was the target of a deepfake video, in which he talks about the theft of information, announced that it will remove manipulated japan mobile database content from the network, which has characteristics of a deepfake.


In addition, a way to improve the detection of deepfakes is also being studied. If a piece of content is classified as false or partially false by one of its fact-checkers, the post will have its reach reduced and, when shared, the user will receive a warning that it is a fake publication. Other technology-focused companies are also looking into ways to eliminate this type of content. Google, Microsoft, among others, have launched the Deepfake Detection Challenge, a challenge to create technology capable of detecting deepfakes.


For the average citizen, some tips can help when assessing whether a video is real or not. Crops of recording locations or even speeches, with changes in speed, in order to create a new narrative, can be considered characteristics of a potential deepfake, as well as blurring and out-of-focus faces, especially around the mouth, where there may be desynchronization of audio and video, facial expressions, voice timbre and intonation. Experts point out that one of the main points to pay attention to is noticing whether the eyes of the person portrayed in the video are fixed. In most deepfake videos, the eyes do not blink, precisely because they are an object that is very difficult to edit.
Post Reply