Australia is preparing to ban anyone under 16 from using social media in December, but a new government-commissioned report shows the technology meant to enforce it may not work equally well for everyone.
The study, released Monday, looked at selfie-based software that estimates a user’s age from their photo. It found the tools are generally accurate, quick, and designed to protect privacy. But the system struggles around the 16-year age cutoff, with accuracy dropping for certain groups.
For example, white users who are a few years older than 16 are unlikely to run into issues. But the report warns that teenage girls and non-white users could be flagged more often as underage. That could mean extra hurdles like providing an ID or parental consent.
Platforms like Instagram and YouTube will soon be legally required to block under-16 users or face fines of up to A$49.5 million ($32 million). The government says the trial showed strong results for people over 19, but teens in the “grey zone” around the cutoff were harder to classify. A 16-year-old, for instance, has an 8.5% chance of being wrongly labeled too young.
Researchers also found lower accuracy for older adults and for users whose appearance is harder for the system to read near policy thresholds.
That raises doubts about whether the ban can be effectively enforced in just three months. “It seems like there are a lot of variations in accuracy,” said Justine Humphry, a media researcher at the University of Sydney. “That variation is concerning, with the tight schedule for the introduction of a system that will need to be robust and working by the end of this year.”
Communications Minister Anika Wells defended the trial, saying it showed there are “many effective options” to verify age while still protecting privacy, even if no single method works for everyone.
