Age-based bans & social media: what Australia’s report teaches us

Australia’s first report on the social media ban has several key lessons for European countries contemplating a similar approach – and for the EU’s special panel on online child safety:

– Based on a survey of 898 parents of children from 8 to 15: “Of the parents who reported their child had an account on each platform prior to 10 December 2025, around 7 in 10 reported that their child still had an account on Facebook (63.6%), Instagram (69.1%), Snapchat (69.4%), and TikTok (69.3%).
=> this is only for those accounts *known to parents*, so the numbers in practice are likely much higher, and these numbers are *despite* significant efforts by social media platforms to deactivate accounts – over 5 million accounts blocked, based on the report.
For context, it’s unlikely that huge numbers under the age of 13 already had accounts, and UNICEF reports there were 5.7 million “under 18s” in Australia in 2023, so it looks like those 5 million cover at least some duplicates.

– “Initial experiences of the SMMA obligation are mixed, with some children aged under 16 appearing to be relieved they are no longer on social media, and others seemingly celebrating their circumvention and retention of accounts.
=> When several of the intended “beneficiaries” of statutory protection celebrate circumvention, “protection” isn’t necessarily the correct framing.

– “The trial found that facial age estimation is a popular and effective form of age assurance – in particular, for confirming that a person is considerably younger or older than a given age threshold. However, facial age estimation is known to have higher error rates for children near the age threshold of 16 years.
=> Age assurance techniques are difficult to get right in the absence of a government-issued ID. Given the privacy and data protection implications, do governments want to force all platforms to process *more* data (with increased fears of governmental surveillance) to avoid false positives/false negatives?

– The report also includes 4 key “observations”, as well as “good practices” and “poor practices” in relation to each. Take these with a grain of salt, as their wording is sometimes a little accusatory, but they provide for interesting reading at least.
In addition, some parts on “poor practices” appear to assume that slapping age assurance on top of any platform that has evolved organically over the years is a straightforward technical fix. Integration of centralised, common and consistent age assurance measures across a wide range of different points of a user journey is not in all cases a simple extra call to an API. I hope they don’t focus their enforcement on those aspects.

Report: https://lnkd.in/e7KhhBuv

I have already repeatedly highlighted some of the arguments against a ban (see https://lnkd.in/egChT7SQ ) but I find the comments of Michael O’Flaherty, Council of Europe Commissioner for Human Rights, to be really worth the read. See notably https://lnkd.in/eX9b3FE6