Meta is bringing back facial recognition with new safety features for Facebook and Instagram
Meta is reintroducing facial recognition tech on its apps, three years after shutting down Facebook’s “face recognition” system due to backlash. The social network will now deploy facial recognition tools on Facebook and Instagram to fight scams and help users who have lost access to their accounts, the company shared in an update.
The first test will use facial recognition to detect scam ads that use the faces of celebrities and public figures. “If our systems suspect an ad may be a scam with a public figure’s image at risk for celeb-bait, we’ll use facial recognition 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.”
The company said it has already started rolling out the feature to a small group of celebs and public figures. It plans to automatically enroll more people “in the coming weeks,” though individuals can opt out of the protection. While Meta already reviews ads for scams, it isn’t always able to catch “celeb-bait” ads, as many legitimate companies use celebrities to promote their products. Monika Bickert, VP of content policy at Meta, said in a briefing, “This is a real-time process. It’s faster and more accurate than manual review.”
Separately, Meta is testing facial recognition to solve a long-standing issue on Facebook and Instagram: account recovery. The company is experimenting with a “video selfie” option that lets users upload a clip of themselves. Meta will then match it to their profile photos when users are locked out of their accounts. It will also be used in cases of suspected account compromise, preventing hackers from accessing accounts with stolen credentials.
The tool won’t help everyone who loses access to a Facebook or Instagram account. Many business pages, for instance, don’t have a profile photo of a person, so those users will need to use Meta’s existing account recovery methods. But Bickert says the new process will make it much harder for bad actors to abuse the company’s support tools. “It will be a much higher level of difficulty for them to bypass our systems,” Bickert said.
With both new features, Meta says it will “immediately delete” facial data used for comparisons, and the scans won’t be repurposed. The company is also making the features optional, though celebrities will need to opt out of the scam ad protection rather than opt-in.
This may spark criticism from privacy advocates, especially given Meta’s troubled history with facial recognition. The company previously used the tech to power automatic photo-tagging, allowing the platform to automatically recognize faces in photos and videos. That feature was discontinued in 2021, with Meta deleting the facial data of more than 1 billion people, citing “growing societal concerns.” Meta also faces lawsuits, notably from Texas and Illinois, over its use of the technology. The company paid $650 million to settle a lawsuit related to Illinois law and $1.4 billion to resolve a similar suit in Texas.
It’s important to note that the new tools won’t be available in Illinois or Texas to start. They also won’t roll out in the United Kingdom or European Union, as Meta is still in talks with regulators there. However, the company hopes to scale the technology globally by 2025, according to a Meta spokesperson.
The first test will use facial recognition to detect scam ads that use the faces of celebrities and public figures. “If our systems suspect an ad may be a scam with a public figure’s image at risk for celeb-bait, we’ll use facial recognition 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.”
The company said it has already started rolling out the feature to a small group of celebs and public figures. It plans to automatically enroll more people “in the coming weeks,” though individuals can opt out of the protection. While Meta already reviews ads for scams, it isn’t always able to catch “celeb-bait” ads, as many legitimate companies use celebrities to promote their products. Monika Bickert, VP of content policy at Meta, said in a briefing, “This is a real-time process. It’s faster and more accurate than manual review.”
Separately, Meta is testing facial recognition to solve a long-standing issue on Facebook and Instagram: account recovery. The company is experimenting with a “video selfie” option that lets users upload a clip of themselves. Meta will then match it to their profile photos when users are locked out of their accounts. It will also be used in cases of suspected account compromise, preventing hackers from accessing accounts with stolen credentials.
The tool won’t help everyone who loses access to a Facebook or Instagram account. Many business pages, for instance, don’t have a profile photo of a person, so those users will need to use Meta’s existing account recovery methods. But Bickert says the new process will make it much harder for bad actors to abuse the company’s support tools. “It will be a much higher level of difficulty for them to bypass our systems,” Bickert said.
With both new features, Meta says it will “immediately delete” facial data used for comparisons, and the scans won’t be repurposed. The company is also making the features optional, though celebrities will need to opt out of the scam ad protection rather than opt-in.
This may spark criticism from privacy advocates, especially given Meta’s troubled history with facial recognition. The company previously used the tech to power automatic photo-tagging, allowing the platform to automatically recognize faces in photos and videos. That feature was discontinued in 2021, with Meta deleting the facial data of more than 1 billion people, citing “growing societal concerns.” Meta also faces lawsuits, notably from Texas and Illinois, over its use of the technology. The company paid $650 million to settle a lawsuit related to Illinois law and $1.4 billion to resolve a similar suit in Texas.
It’s important to note that the new tools won’t be available in Illinois or Texas to start. They also won’t roll out in the United Kingdom or European Union, as Meta is still in talks with regulators there. However, the company hopes to scale the technology globally by 2025, according to a Meta spokesperson.