On social networks, some users create accounts that try to impersonate another person. The goal? Mislead users.
Falsified companies and profiles are common targets for malicious individuals. And to make matters worse, more and more tools allow them to perfect the creation of fake people and therefore, to reinforce their scam.
To solve the problem of false images or profile photos, the company V7 Labs has created an extension for Chrome capable of detecting if the profile photo is that of a real person or if it is the result of an artificial intelligence.
“There is a lot of misleading content on social media these days and one of its propagators are fake profiles”said Alberto Rizzoli, co-founder of the company, in a video posted on Loom.
How it works ?
Software company V7 Labs has created a new Google Chrome extension that can detect fake profile photos created by artificial intelligence (AI).
Clearly, the company has developed “Fake Profile Detector”, a new, free tool. It is a browser extension available on Google Chrome. It would be able to identify a fake profile picture or image with 99.28% accuracy.
To start the detection process, the user just needs to right-click on an image. Then he will be able to select the option “Check fake profile picture” and detect if the picture is a Deepfake (GAN) or a real person. In a final step, the artificial intelligence provides the user with a percentage estimate. And that in just a few seconds. A video offers a demonstration of this extension.
Stemming misinformation online
The software would also be able to deal with fake news and misinformation. “This Chrome extension will help authorities and even ordinary people spot and report profiles that spread fake news and create misinformation. And even ordinary people to spot and report profiles that spread fake news and create information chaos, especially in the face of the Russian invasion of Ukraine”concludes Alberto Rizzoli, co-founder of V7 Labs with PetaPixel.
It is possible to add this AI software to your Google Chrome via this link.
What level of efficiency?
The co-founder clarifies that, for the moment, the software is still unable to detect high-quality “deepfakes” that are presented in videos. In the same vein, the lemon-press site reports that its team tested the tool on certain Twitter profiles generated via the ThisPersonDoesntExist portal. “Unfortunately, the AI got it wrong twice”, regret the site. He adds, “so you have to be aware that it is not 100% reliable”. In addition, according to ghacks.net, the reliability rate is higher on deepfakes generated via the Bored Humans site.
Note that the software in question uses neural networks to modify images in an extremely realistic way. The AI neural network is a group of mathematical algorithms producing output data from input data. We then speak of input and output. These neural networks are trained using a multitude of input data coupled with their respective output data. Once the output data has been calculated, they compare it to the actual known output data. In other words, it compares the generated fake profile picture with a profile picture of a real person. These software are constantly updated to improve the results and make them more and more undetectable.
In addition to this extension, it is still possible to detect certain profiles generated with the AI with the naked eye. This is because a neural network sometimes makes mistakes, hence the appearance of artifacts such as misfolded pattern, strange hair color, etc. Thus, some clues of the fake can be found in the hair, eyes, or even the neck.
We would like to say thanks to the writer of this article for this remarkable web content
A tool to spot fake online profiles – Geeko
We have our social media profiles here as well as other pages related to them here.https://www.ai-magazine.com/related-pages/