Facebook has implemented new updates aimed at combating deepfakes on its platform, particularly focused on stopping fraudulent ads that use celebrity images to deceive users. This tactic known as 'celeb-bait' has posed a recurring challenge for the social network, where scammers create seemingly legitimate ads with public figures to redirect to fraudulent websites and generate clicks.
Facebook's Vice President of Content Policy, Monika Bickert, explained that the company will implement facial recognition technology to analyze images of celebrities in ads. This system will use artificial intelligence to compare photos of public figures in the ads with their profile images on Facebook and Instagram to determine possible attempts at deception.
In a roundtable with media, Monika Bickert detailed how the system will work: 'The software will perform a facial assessment. It will look at the features of a face and then compare them with known images on the celebrity's Facebook or Instagram accounts.' Although they acknowledge that well-made deepfakes could bypass this technology, they are confident that facial recognition will significantly improve the detection and removal of these frauds.
One of the existing concerns is whether AI-generated deepfakes could deceive the facial recognition system. Bickert emphasized that while there is a possibility that extremely realistic deepfakes may evade this measure, they are confident that the system will improve over time to identify such content.
The implementation of this technology is part of a broader effort by Facebook to protect users and celebrities from scams on the platform. Besides fraudulent ads, the social network will seek to combat the creation of fake accounts of public figures by using facial recognition to identify and eliminate these counterfeit accounts more quickly, thereby ensuring the safety of public figures and their followers.
