Instagram’s process for determining whether a user is over 16 is relatively quick and painless if you’re clearly an adult – but how does it work if a 13-year-old tries to change their account’s date of birth to make them appear grown up?

Meta in November began notifying Instagram and Facebook users whose date of birth is set as under 16 – or who the platform understands to be under 16 – that their accounts will be deactivated as part of Australia’s social media ban for children. The ban takes effect on 10 December, but Meta has said it will start removing access to users under 16 from 4 December.

_{Sign up: AU Breaking News email}

As part of Guardian Australia’s reporting on what the platforms show to various age demographics, a phone was set up with dummy social media accounts.

One was set up on Instagram, with an age set to 15, to see what would happen when the under-16s social media ban came into effect. Instagram subsequently sent a notification: “Due to laws in Australia, soon you won’t be able to use social media until you’ve turned 16.

“You will not be able to use your Instagram account until you’ve turned 16. This means you can’t use Instagram and your profile won’t be visible to you or others until then.

“We’ll let you know when you can use Instagram again.”

The account was then presented with two options: download the account data and prepare for it to be deactivated until the user turns 16, or review the date of birth.

Choosing the latter allows the user to take a “video selfie” to prove the account holder is over 16. The app accessed the front-facing camera and required the test user, an adult who has thick facial hair, to move their head from side to side, similar to the verification method used to set up face-unlock on a smartphone.

A notification then stated it usually took between one and two minutes to verify, but could take up to 48 hours.

The app quickly stated the account, set up by the test adult user, was marked as over 16.

In a separate test, a 13-year-old set up a new account on a phone that had never had Instagram installed using a date of birth clearly showing they were under 16. There was no immediate notification of the looming social media ban.

When the child then attempted to change their date of birth to one of an adult, the same video selfie facial age estimation process was undertaken.

Within a minute, it stated “we couldn’t confirm your age” and then requested government ID to confirm the user’s date of birth.

Facial age testing in the age-assurance trial data showed that people over 21 would generally be much less likely to have an issue with being declared over 16. Those closer to the age of 16, as well as minority groups, were shown to have a higher false positive or false negative rate.

Meta may have already assessed users who have yet to receive a notification as being over 18 based on information such as date of birth, the length of time a person has had an account, and other account behaviour activity.

A Meta spokesperson said the experiment showed the process works as intended, “with the adult user being able to verify his age and move on, and the under 16 user being age checked when they tried to change their age.”

“That said, we must also acknowledge the findings of the Age Assurance Technology Trial, which recognises the particular challenges of age assurance at the 16-age boundary, and we anticipate that at times the process may not be perfect,” the spokesperson said.

Last month, the communications minister, Anika Wells, acknowledged there would be teething issues as the ban comes into effect.

“We know this law will not be perfect, but it is too important not to have a crack,” she said.

Meta uses Yoti for its age assurance. The company states on its website that facial images are deleted after the check is completed.

The ban will affect Meta’s Facebook, Instagram and Threads platforms, along with Kick, Reddit, Snapchat, TikTok, Twitch, X and YouTube.