Meta(META.US) restarts facial recognition technology to target "celebrity scam" fraud

Meta Platforms Inc. (META.US) is using facial recognition technology to combat ads that use celebrity images to defraud people, known as "celebrity bait ads." By comparing the images in posts to those of celebrities on their Facebook or Instagram accounts, Meta can identify and block ads that try to increase credibility by using celebrity photos. The company said in a blog post that if a match is confirmed and the ad is determined to be fraudulent, it will be blocked. Meta did not disclose how common such fraud is on its platform. Meta's move is part of its efforts to combat spam and rule-breaking content. The company has nearly 3.3 billion daily active users and relies on artificial intelligence to enforce content rules and guidelines to handle the large number of daily rule-breaking reports. However, the automated systems' errors have also led to some legitimate accounts being accidentally suspended or blocked. In addition, Meta plans to use facial recognition technology to help users whose accounts have been locked. In a new test, users can submit a selfie video, which Meta will compare to the photos on the account to confirm identity. The process takes only a minute, compared to the more cumbersome traditional ID or official certificate verification. Meta promises to delete all generated facial data immediately, regardless of whether the match is successful. Notably, Meta has a complex history with facial recognition technology. The company has faced lawsuits in multiple states in the US for using the technology without users' consent, and was ordered to pay $1.4 billion in 2024 to Texas. In addition, it paid $6.5 billion in another lawsuit in Illinois. Monica Bickert, Meta's content policy vice president, said the company will not test video selfie verification in Illinois or Texas. Meta's move is part of its efforts to combat spam and rule-breaking content. The company has nearly 3.3 billion daily active users and relies on artificial intelligence to enforce content rules and guidelines to handle the large number of daily rule-breaking reports. However, the automated systems' errors have also led to some legitimate accounts being accidentally suspended or blocked. In addition, Meta plans to use facial recognition technology to help users whose accounts have been locked. In a new test, users can submit a selfie video, which Meta will compare to the photos on the account to confirm identity. The process takes only a minute, compared to the more cumbersome traditional ID or official certificate verification. Meta promises to delete all generated facial data immediately, regardless of whether the match is successful. Notably, Meta has a complex history with facial recognition technology. The company has faced lawsuits in multiple states in the US for using the technology without users' consent, and was ordered to pay $1.4 billion in 2024 to Texas. In addition, it paid $6.5 billion in another lawsuit in Illinois. Monica Bickert, Meta's content policy vice president, said the company will not test video selfie verification in Illinois or Texas.