Meta launches technology to detect fraudulent advertisements with “celebrities” and simplifies account recovery

Meta✴ has announced that it is testing facial recognition technology to combat fraudulent advertisements that feature images of celebrities. Facebook✴ and Instagram✴ will also use the new feature to identify fake accounts of famous people created to deceive users.

Image source: Gerd Altmann/Pixabay

As TechCrunch explains, the purpose of these tests is to enhance existing anti-fraud measures, such as machine learning-based automated scans that check ads on Facebook✴ and Instagram✴. If an ad is flagged as suspicious, the identification process will immediately begin with account profile photos.

«

Meta✴ states that facial recognition technology is used solely to combat fraudulent ads and once verified, all images, whether a match is found or not, are deleted and not used for any other purpose. Initial tests involving a small group of public figures showed “encouraging” results regarding the speed and effectiveness of identifying and blocking fraudulent advertising. The company is also considering using the technology to detect deepfake ads created using generative AI, and is testing facial recognition to detect fake celebrity accounts.

In the coming weeks, the company will begin sending in-app notifications to a wider range of public figures whose images were used in fraudulent ads, informing them that they are automatically included in the protection system. However, according to Bickert, you can opt out of this feature at any time through the Account Center.

Among other things, Meta✴ has begun testing video selfies using facial recognition to speed up the process of restoring access to accounts that have been blocked for any reason. This method will make the process of restoring access faster and easier than uploading an ID image. It is noted that “video selfie verification expands the possibilities for restoring access to your account, takes only a minute and is the easiest way to confirm your identity.” This method is similar to identification using Face ID on iPhone.