How Meta is using AI to identify teen accounts

SALT LAKE CITY — Can technology really tell how old someone is online? According to a Meta spokesperson new AI-powered tools are helping identify underage accounts and place teens into age-appropriate experiences.

Social media platforms have age restrictions, but verifying a user’s real age has always been a challenge. Most platforms require users to be at least 13, yet that system relies heavily on people entering their own birthdate.

“That’s really what we’ve been trying to solve for and now what we’re talking about is some advancements we’ve made our ability to use artificial intelligence and technology advancements to really help strengthen those efforts, to really help us enforce against out age restrictions,” said Jennifer Hanley, Meta head of safety. 

Meta said new AI-powered age assurance tools help identify accounts that may belong to underage users.

“We’re now able to use AI to really help us find and remove accounts that may belong to people that are under the age of 13,” said Hanley. “In the past we’ve been able to look at text and information, someone may be admitting that they’re 12 years old. Now what we’re able to do is look at a whole variety of signals. So, it may be someone posting about their birthday or what grade they’re in. And we’re also able to look across different surfaces that we have.” 

Those signals can include activity across profiles, posts, comments, and videos.

“We’re also starting to use visual analysis so visual cues, that might be related to someone’s height or facial structure. Other things indicate they’re a child so we can also work to deactivate and remove them as well,” said Hanley.

Meta said the technology is also designed to make sure teens aren’t placed into adult experiences. Teens misrepresenting that they are an adult may never have been put into a teen account.

“There might be certain circumstances where teens are misrepresenting that they are an adult. But they’re a teen and should never put into our teen account experience,” said Hanley.

Those teen account settings include built-in safeguards for younger users.

“They include privacy protections, restrictions on messaging, and who can message a teen. We have time limits. We even have a new content experience– really meant to be age-appropriate content for teens,” said Hanley.

When an account is flagged, users may be asked to verify their age.

“If we detect that someone might be misrepresenting their age, we’ll put them into the teen account settings and give them an opportunity to prove their age,” said Hanley. “They can age-verify either using an ID or we have a partner called Yoti that also helps us with facial age estimation as well.”

Meta said age verification is not something any one company can solve alone.

“This is an ecosystem-wide issue. Everyone is trying to figure out how to best understand the age of who’s on their platform,” said Hanley. 

The company believes centralized systems could help standardize age assurance.

“We think that governments can play a really important role by passing legislation that calls for age verification and parental approval at the App Store level,” said Hanley.

For parents and educators looking for more information, Meta said educational resources are available.

“Meta has a family center with a lot of great information for parents, for educators, for everyone. We have educational information resources, conversation starters. We have information about our teen accounts and even parental supervision tools,” said Hanley.

Meta said improving age assurance helps keep teens in experiences designed for their safety, while giving parents more confidence in how platforms protect younger users.

“Utah has taken an important step when it comes to understanding age, and was one of the first states to pass App Store legislation. So I think that Utah really understands the importance of having this framework in pace,” said Hanley.

 

Related: Instagram says it’s safeguarding teens by limiting them to PG-13 content