Meta is reintroducing facial recognition technology into its platforms, over three years after discontinuing Facebook’s “face recognition” feature amid widespread concerns about the technology. The company aims to utilize facial recognition on Facebook and Instagram to combat fraudulent activities and assist users in regaining access to their accounts, as stated in a recent update.
The initial phase of testing will focus on identifying fraudulent advertisements that exploit the images of celebrities and other public figures. “If our systems suspect that an ad may be a scam that contains the image of a public figure at risk for celeb-bait, we will try to use facial recognition technology to compare faces in the ad against the public figure’s Facebook and Instagram profile pictures,” Meta explained in a blog post. “If we confirm a match and that the ad is a scam, we’ll block it.”
This feature has already been introduced to a select group of celebrities and public figures, with plans to automatically include more individuals in the upcoming weeks. However, there is an option for these figures to opt out of this protection.
Meta’s existing ad review systems are not always successful in detecting “celeb-bait” ads, as many legitimate businesses also use celebrities in their marketing. Monika Bickert, Meta’s VP of content policy, explained in a briefing that the new facial recognition feature is a real-time process that is both quicker and more precise than manual reviews.
In a separate initiative, Meta is experimenting with facial recognition for account recovery on Facebook and Instagram. The company is testing a “video selfie” feature that matches a user’s video clip to their profile photos to assist with account access issues, particularly when an account is suspected of being compromised. This tool is designed to prevent unauthorized access by hackers using stolen credentials.
The “video selfie” recovery option will not be applicable to all account types; for instance, business pages without personal profile photos will need to rely on Meta’s existing account recovery methods. Bickert asserts that this new process will pose a significant challenge for malicious individuals attempting to circumvent Meta’s security measures.
Meta assures that any facial data used for these new features will be “immediately deleted” after the comparison and will not be repurposed. Both features are optional, with celebrities needing to opt out of the scam ad protection rather than opting in.
Privacy advocates may express concerns, especially considering Meta’s complex past with facial recognition. The technology was previously employed for automatic photo-tagging, enabling the company to identify users’ faces in photos and videos automatically. This feature was discontinued in 2021, and Meta deleted the facial data of over 1 billion users, referencing “growing societal concerns.”
Meta has also faced legal challenges, including lawsuits in Texas and Illinois, over its use of facial recognition. The company settled an Illinois lawsuit for $650 million and resolved a Texas case for $1.4 billion. Notably, the new facial recognition tools will not be available in Illinois or Texas initially, nor in the United Kingdom or European Union, as Meta continues discussions with regulators in those regions.
However, Meta is optimistic about expanding this technology globally by 2025, according to a company spokesperson.
