Australia’s under-16 social media ban faces doubts as teens outsmart AI checks

The trial, conducted by tech firm KJR, reviewed nearly 60 tools, concluding facial recognition was relatively fast and accurate, whereas methods involving credit cards or gesture inputs lacked precision and practicality for borderline age groups

As Australia prepares to roll out a landmark nationwide ban on social media use for children under 16, a government-backed trial of AI-driven age-verification software has revealed both promise and pitfalls in enforcing the new regulation.

Set to take effect from December, the legislation will require platforms like Instagram, Facebook, TikTok, and Snapchat to take “reasonable steps” to block access for users under 16, or face penalties of up to A$49.5 million. While the goal is to safeguard young users from online harms such as cyberbullying, body image issues, and toxic content, the implementation has sparked concerns over its effectiveness and enforceability.

According to a Reuters report, as part of a recent pilot, around 30 students, including 13-year-old Jasmine Elkin from Perth, tested five facial recognition-based age-estimation tools. Despite being impressed by their accuracy—some tools could determine age down to the month—Elkin remains unconvinced about their ability to prevent determined teens from bypassing the system. “They can just ask a sibling to pose for the photo,” she said. “People will always find a way.”

The trial, run by tech contractor KJR, was designed to evaluate the reliability of nearly 60 proposed solutions. General Manager Andrew Hammond noted that while the facial recognition methods proved fast and fairly accurate, other tools—like those requiring credit card details or complex hand gestures—were less practical or too imprecise, especially near the 16-year-old threshold.

Tech-savvy youth find loopholes

As per the report, for teenagers like 14-year-old Charlie Price from Canberra, the test was a preview of life without social media. “I use it a lot, but I can live without it,” he said, though he acknowledged some peers may be upset by the ban. Price, along with others, noted that many teens would likely look for workarounds, including switching platforms or faking their age.

The software trial also demonstrated how tech-savvy today's youth are. Testers moved through systems so efficiently that organisers had to double the number of tools trialled and reduce testing times. Despite this, some tools produced wildly inaccurate results, such as one case where a 13-year-old was estimated to be 42.

Educators like Nathanael Edwards, principal of Radiant Life College in Queensland, echoed these concerns. His students tried a basic age-gate using birthdates—some succeeded in faking older ages, while others were tripped up by simple math errors.

The government is expected to receive a full report by the end of July, which will inform the eSafety Commissioner’s final recommendations. Until then, the key question remains: How accurate is accurate enough?

“There's no agreed benchmark yet,” said Hammond. “Is 70% good enough, or does it have to be 100%? That’s something the government still needs to decide.”

Also Read: Australia Imposes Strict Age Limits on Social Media Use