Australia Allows Flexibility as Age-Checks Prove Imperfect for Under-16 Social Media Ban

Australian legislation banning under-16s from holding social media accounts will permit platforms to make some errors when checking age, government and watchdog officials have clarified, as new trials show limits in technology and equity in age verification.

The law, passed in November 2024, comes into force on 10 December 2025, obliging “age-restricted social media platforms” to take “reasonable steps” to stop users under 16 from creating or retaining accounts. Platforms may face fines up to AUD 49.5 million should they fail to comply.

An extensive age assurance trial, involving 60 technologies from 48 vendors, finds that though many tools are effective, none are perfect. Facial age estimation, phone-operating-system checks, parental consent, and ID verification all have trade-offs. Errors cluster especially among users close to the 16-year threshold, and among Indigenous Australians, non-Caucasian users, females, and others underrepresented in training data.

Government guidance mandates that platforms need not test the age of every user or achieve a specific accuracy rate to avoid penalties. Instead they must show “reasonable steps” in their age screening, be transparent about their methods, offer multiple verification options, and maintain appeals or error-correction mechanisms for users falsely excluded or included.

Privacy is central to the architecture of the law. Platforms may use less invasive inference based on behaviour or metadata, or age‐estimation technologies. While use of government-issued ID is permitted, it cannot be the sole accepted method. Platforms are required to offer alternatives. Users’ age data must not be automatically repurposed, and protections around retention of data are being emphasised.

The trials show that the average facial age estimation takes under 40 seconds. But in many cases for users aged 14-17, the process may take far longer, especially when fallback verification is required. Some providers showed error rates as high as 34 per cent for 14-year-olds and 73 per cent for 15-year-olds being incorrectly assessed as 16 or older.

Platforms such as Facebook, Instagram, Snapchat, TikTok, X, Reddit, and YouTube are within scope. Some services such as messaging-only apps or education/health platforms are exempt. Under the law, even existing accounts held by under-16s must be located and deactivated.

Officials acknowledge the law will not be fully effective immediately. Early efforts will focus on removing underage accounts already existing and setting up proper mechanisms, while preventing new underage registrations will require more robust technical infrastructure. Platforms have until December to demonstrate compliance, but with recognition that perfection is not expected.