SYDNEY/NEW YORK — Three years after discontinuing facial recognition software on Facebook due to privacy concerns and regulatory scrutiny, Meta announced on Tuesday (Oct 22) that it is testing the technology once again to combat "celeb bait" scams.

Meta plans to enroll around 50,000 public figures in a trial that will automatically compare their Facebook profile photos with images used in suspected scam ads. If a match is found and the ad is deemed fraudulent, Meta will block it.

The enrolled celebrities will be notified and given the option to opt-out if they choose not to participate, the company said.

The trial is set to be rolled out globally starting in December, but will exclude certain major regions, including the UK, the European Union, South Korea, and US states like Texas and Illinois, where regulatory approval has not been granted.

**Related:**

Meta faces US state lawsuits over teen social media addiction.

Meta’s vice president of content policy, Monika Bickert, said the initiative focuses on protecting public figures whose likenesses are frequently used in scam advertisements.

"The goal is to provide as much protection as possible for them. They can opt out if they prefer, but we want to make it easy for them to access this protection," Bickert said in a press briefing.

This test highlights Meta’s effort to balance the use of controversial technology with the growing regulatory concerns over scams, while addressing the longstanding issue of user data handling that social media companies have faced.

In 2021, Meta deactivated its facial recognition system and deleted the facial scan data of one billion users, citing "increasing societal concerns." In August, the company was ordered to pay Texas US$1.4 billion (S$1.8 billion) for allegedly collecting biometric data illegally.

Simultaneously, Meta is dealing with lawsuits accusing it of failing to curb "celeb bait" scams, which use AI-generated images of celebrities to defraud users into investing in fake schemes.

Under the new trial, any facial data generated during comparisons with suspected scam ads will be deleted immediately, regardless of whether a scam is detected.

Meta stated that the tool has undergone its internal "robust privacy and risk review process" and was discussed with regulators, policymakers, and privacy experts before the testing phase began.

The company also mentioned plans to test facial recognition as a method for non-celebrity users on Facebook and Instagram to regain access to accounts compromised by hackers or locked due to forgotten passwords.

Disclaimer: This image is taken from Reuters