Meta is testing a new facial recognition tool across Facebook and Instagram to detect scams involving celeb baits. According to reports, 50,000 images of public figures will be used in the test and will expand globally by December prioritizing the celebrities whose images are already involved in the scam.
As a part of eliminating the fake ads from the platform, Meta is testing the facial recognition service across Facebook three years after the initial shutdown of the facial recognition software due to the existing regulatory and privacy limitations set by the US.
Meta on Tuesday said it was testing a new facial recognition technology to control and eliminate the mishandling of celebrity photos and videos as part of fraud or impersonation on social media platforms including Facebook and Instagram.
This new technology aims to prevent ‘celeb baits’ which use photos and videos of celebrities to create ads and scam users thereby convincing them to share personal information or send money. Many users think the celebrities performing in the ads are real and legitimate, and send money or share information blindly.
Since these ads look like real promotions, they are difficult to detect and eliminate. To eliminate these scams, Meta uses the new facial recognition technology in their ad review system. When an ad is flagged and detected with a celebrity image, the system compares the photo in the ad with the celebrity’s images in their official Facebook or Instagram account. If both images match and the ad turns out to be a scam, Meta blocks or removes the ad without any prior notice.
The company assures that the facial data collected for the comparison is permanently removed from the server data right after the comparison, regardless of whether the system finds a match. The facial verification technology is similar to identity verification used to unlock a mobile device or other apps.
Meta has been working with this facial verification system using a smaller range of celebrities and the initial stage was a huge success. The company is planning to expand its trial with around 50,000 images of public figures around the globe, excluding some strict jurisdictions including Britain, the European Union, South Korea, the US states of Illinois and Texas.
The celebrities will be notified about their enrollment and they have the choice to stay in or opt out if they don’t want to be a part. The company is focusing on selected celebrities whose images have been identified as used in scams to increase the effectiveness of the recognition system, said Monica Bickert, Meta’s vice president of content policy on an interaction with journalists.
Currently, Meta is using a combination of automated systems and reports from users to detect the scams in ads. The implementation of facial recognition technology adds an additional layer of security and eliminates the occurrence of celebrity Ad scams before they reach the audience to a wider extent.
Meta is also planning a facial recognition method to help users access their Facebook or Instagram accounts once they are locked or hacked by using a video selfie. But as of now, the company is asking to upload official documents for verification.
In the video selfie recognition method, a user has to upload a video selfie which is then compared with the user profile picture if both match, the account access is sanctioned, declined otherwise.
The uploaded video selfie is encrypted and stored securely in their servers and is removed immediately after the comparison process is completed. This method is comparatively faster and more secure than document verification.
Earlier in 2021, Meta shut down the facial recognition system, deleting the face recognition data of over 2 billion users. In August this year, Meta was ordered to pay $1.4 billion to settle the case of collecting biometric information of the users illegally. This could be the main reason for the more emphasized privacy policies in the new face recognition tests.
It’s natural that scammers find ways to intrude or break the safety measures, however, the regular and informed security measures taken by Meta help the platform stay ahead of the scams and keep their users safe. Meta also values the privacy of the users as it notifies the users prior to enrollment and removes the data right after the comparison.
Latest Tech News: Huawei Releases Homegrown HarmonyOS NEXT Operating System
