Why can’t Instagram verify users’ ages and keep kids off the app?

At Instagram’s virtual creator conference last summer, a conversation between the app’s chief executive Adam Mosseri and the all-singing, all-dancing teen influencer JoJo Siwa took an awkward turn.

Eighteen-year-old Siwa, known for wearing bubblegum colours and hair bows, had begun to outline her years-long journey to social media stardom.

“I’ve had everything since I was five! I know you’re not supposed to have Instagram until you’re 13. I did, I had an account. Many five year olds do—”

Swooping in to cut Ms Siwa off mid‑sentence, Mr Mosseri said, “I don’t wanna hear that.” He was chuckling, but the exchange was painful to watch.

Here was an ambassador for the photo app inadvertently betraying that it has long been swarming with underage users, a breach of its own rules.

And, in response, the top executive effectively covering his ears.

Under-13s are not allowed on Instagram, which is owned by Meta, partly to comply with United States privacy laws.

Nevertheless, a survey of more than 2,000 minors published last year by the children’s charity Thorn found that 65 per cent of nine- to 12-year-olds have used Instagram at least once, and 40 per cent use it at least once a day.
Those who do are likely to be exposed to numerous harms.

According to Thorn, some 38 per cent of girls and 36 per cent of boys between the ages of nine and 12 say they have been bullied or made to feel uncomfortable online; 14 per cent have been asked to send a nude video or photo by someone online.

Instagram collects and crunches reams of personal data in order to profile users and serve up tailored advertisements. It is now so good at guessing consumer desires that it has been forced to stress — repeatedly — that it isn’t hijacking mobile phones to listen to private conversations.

How then can the app with the US$660 billion parent company know that I want new red Adidas trainers, but not know my age?

In a blog post last July, Facebook maintained that age is difficult to assess and that the technology for both Facebook and Instagram was a “work in progress”.

It was, it said, building artificial intelligence to proactively find under-13s who have lied about their age — by scanning for what age people mention when they post wishing you a happy birthday for example — but added that the technology “isn’t perfect”.

It also said it was “developing a menu of options for someone to prove their age” without having to ask for ID, though the company has yet to unveil those features. Instagram’s critics argue that it has deliberately not done enough to eject underagers. User numbers look plumper for shareholders if they remain.

Baroness Beeban Kidron, chair of the children’s digital rights charity 5Rights, believes the idea that this is too great a technological challenge is misleading.

“It’s all because it’s not in their commercial interest to do so,” she says.

Indeed, there is now a growing market of solutions in the US$1 billion “safety tech” sector, typically wielding machine-learning technologies.

London-based start-up Yoti uses your camera to analyse your face for its “facial age estimation” tool and claims to have a margin of error of around one and a half years for six- to 19-year-olds.

BioCatch, which specialises in helping banks catch fraudulent activity, claims to be able to discern age by analysing the way a person uses their devices.

For example, older users will use one finger to type on their phone, where younger users will use both thumbs, and so on.

Other experts are pushing for more centralised age verification: Should Apple confirm age through ID checks as a way of blocking underage users from downloading inappropriate apps altogether?

This is something that Instagram has gently started to lobby for.

For critics, it is remarkable that Facebook and Instagram waited until now to start to develop its own age verification technology.

Other social media apps, such as TikTok, face similar criticism.

Mary Aiken, a safety tech expert and cyberpsychology professor at Capitol Technology University in the US says, “Until there is a regulatory environment that mandates consequences of not deploying accurate age verification then the situation will remain as it is” — in other words, “hopeless”.

Politicians and watchdogs have begun to circle.

Instagram itself announced plans to build “Instagram Kids” for under-13s, arguing that preteens live on the internet no matter what, so better to carve out a safe space with extra parental controls.

But the initiative generated so much pushback that the company has paused the idea temporarily while executives consult with experts. Which is still a very slight improvement over the CEO looking the other way. FINANCIAL TIMES

ABOUT THE AUTHOR:

Hannah Murphy is a tech correspondent in San Francisco, covering Facebook, social media, cyber security, among other issues.