JAKARTA – Facebook continues to try to eradicate the issue of misinformation circulating on its platform. One of them is by presenting Deepfake detection software.
Cited from Sputnik News, Deepfakes are media that include photos, videos, and audio recordings that are edited using artificial intelligence (AI) so that they look realistic.
This often causes confusion among internet users in general and Facebook social media users in particular. Because they are not able to distinguish between fake and real images or videos. Deepfakes have the potential to distort facts into hoaxes.
One of the computer programmers who works for Facebook explained that his party has developed software that is not only able to recognize Deepfakes, but can also detect the origin of the file, as written on the official Facebook blog.
Facebook Artificial Intelligence (AI) researchers Xi Yin and Tal Hassner teamed up with a team of computer science experts to reveal that they can "undo" Deepfake files through reengineering technology. The technology is able to detect facial images and determine how and where they were created.
"This work will provide researchers and practitioners with the tools to better investigate incidents of organized disinformation using Deepfakes, as well as open up new directions for future research", the scientists said in a statement published by Michigan State University.
SEE ALSO:
Not only that, but the researchers also revealed that the process of creating such Deepfakes usually changes the "fingerprint" of digital files, and leaves certain defects. This Deepfake detection technology will be used to look for defects in Deepfake files.
"In digital photography, fingerprints are used to identify the digital camera used to produce the image", the scientists said.
“Similar to a device fingerprint, an image fingerprint is a unique pattern left on an image… which can both be used to identify the generative model from which the image was created”.
In addition to the Facebook team, similar technology has also been created by Microsoft called Video Authenticator. Microsoft has introduced this technology at the end of 2020 in the midst of the US presidential election.
The technology made by Microsoft is also able to reveal Deepfake files by analyzing images and videos and detecting manipulations that cannot be seen by the human eye.
With the presence of Deepfake detection technology, in the future, it is hoped that in the future it will be able to combat the spread of misinformation in cyberspace, especially image and video files and audio that have been edited with AI that are used to distort facts.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)