Six months out from teen social media ban, age-checking tech mistakes kids for 37-year-olds
https://www.abc.net.au/news/2025-06-19/teen-social-media-ban-technology-concerns/105430458

Tests of automated age-guessing technology to be used for a proposed teen social media ban in Australia revealed it's wildly inaccurate and not fit for purpose.

Today, @daedalus took part in a government briefing and Q&A session, and as you can see from this thread, it was a total shitshow: https://eigenmagic.net/@daedalus/114712755649100993

#AusPol #AgeAssurance #SocialMedia #TeenSocialMediaBan #Privacy #Security #CyberSecurity #Cyber #CyberCyberCyberCyberCyber

Six months out from teen social media ban, age-checking tech mistakes kids for 37-year-olds

Children as young as 15 were repeatedly misidentified as being in their 20s and 30s during government tests of age-checking tools, sowing new doubts about whether the teen social media ban is viable.

ABC News

@daedalus Dunno whether to call this Teen Social Media Ban schemozzle "policy on the run" or if there's some other better fitting name for it, but it sure is a mess.

There's no "one size fits all" solution, and anyone trying to sell one is lying or an ignorant fool. I would go as far as saying there is no totally effectively solution or combination of solutions.

Albo was right to say "Government may not be able to protect every child from every threat on social media." He knows they cannot, yet he's going to try because he thinks it's a good look.

"Teen Social Media Ban" is a feel-good policy to appease voters who say "I'm scared of what's on the Internet", or "Won't somebody think of the children?", or "What are kids up to these days?"

The government has promised to do something about it, so now they have to be seen to do something.

Will they deliver something?
Sure.

Will that "something" work as promised?
Nope.

Will it have unintended consequences?
Yes.

Will it have unforeseen side-effects that hurt innocent people?
Unfortunately.

Will it have security holes that are exploited by child-abusers, criminals, bent cops overreaching their powers, and more?
Yes.

Is it going to be expensive?
OMG, yes. The big Consulting firms are lining-up, and Tech giants know if they complain hard enough, they'll get paid well to have a hand in specifying how it's implemented, get well compensated for the inconvenience of any changes they're required by these new regulations to make at their end, or ideally all of the above.

Will it make the internet a safer place?
That's so difficult to measure that I'll remind you of what I said above about people making such statements being liars or fools.

Will it add another unnecessary layer of complexity their everyday life when using the internet in Australia?
Yes. People will have to accept it as the new "normal", and hand-in-hand with that, greater numbers of people will be adopting VPNs, browser plugins, and other technology to circumvent this ill-advised "Teen Social Media Ban".

#AusPol #AgeAssurance #SocialMedia #TeenSocialMediaBan #Privacy #Security #CyberSecurity #Cyber #CyberCyberCyberCyberCyber

@BinChicken @daedalus Colour me gobsmacked.

Not!

Foolish pollies. Duplicitous techies. Credulous [cohort of the] public.

@BinChicken @daedalus
Stopping kids from using Facebook or Twitter is probably a good idea overall. Even better would be stopping everybody.

The real problem is not kids using social media. It is WHO they end up socialising with that is the problem.

However humans are communicating organisms. They will find some way.

How long before someone puts Mastodon on GNUnet? That would be fun to regulate.

@BinChicken @daedalus The next generation of kids, when stores are run by AI, are going to have so much an easier time getting booze than I did! Lucky punks!

@BinChicken @daedalus More on this abject silliness.

https://www.crikey.com.au/2025/06/20/teen-social-media-ban-trial-all-methods-vpns-parents-help/

QUOTE

Australia’s federal government had a “world-first” idea for how to keep our kids safe online.

Batting away expert concerns about how it would work, the government pushed ahead. It poured time and money into a scheme meant to stop children accessing certain parts of the internet.

This was in 2007, not 2025, back when the Australian government pursued its infamous internet porn filter.

That government was publicly embarrassed by a precocious teen, Tom, who says he was able to bypass the $84 million filter in just half an hour.

Almost two decades later, some of the experts who have been part of testing the methods for enforcing the Albanese government’s planned teen social media ban are worried history is about to repeat itself.

While there are unanswered questions about how well the ban will work in practice — an ABC report said that facial analysis tech tested by the trial could accurately estimate someone’s age within an 18 month range 85% of the time — another major concern is how people might thwart or work around these technologies.

Even before the ban passed parliament, the government said that its measures wouldn’t be foolproof, but it hoped to be as tightly enforced as possible.

“Government may not be able to protect every child from every threat on social media but we do have a responsibility to do everything we can, to help as many young Australians as we can,” Prime Minister Anthony Albanese said.

The law that passed parliament in late 2024 was a barebones document. It started a countdown until the law would pass and set in motion a process to develop the rules of how the ban would work.

Separate, but linked, was a $6.5 million trial commissioned by the government to investigate how a social media minimum age could be enforced. Its findings would inform the “reasonable steps” established by the government that social media companies would have to take when gauging a user’s age in order to enforce the teen social media ban.

The Age Assurance Technology Trial’s winning tenderer was a coalition led by UK company Age Check Certification Scheme (ACCS). The coalition would be responsible for assessing “age assurance technologies” — like digital ID, facial analysis and other novel methods of figuring out someone’s age online — for “effectiveness, maturity, and readiness for use in the Australian context”, and publishing a report on its findings.

The ACCS project plan, written in November before the law was passed or the tender was publicly awarded, said the group would test the technologies for detecting fake documents, deepfaked video and other security exploits.

Several months later, after the law had been passed and the tender awarded, the ACCS published an evaluation proposal plan that laid out which “circumvention” methods would and wouldn’t be tested.

It said the trial would test if the technology could identify a person in a disguise or using a photograph of someone, but that it would not test for ways that people might “make deliberate, concerted efforts to evade the age assurance check which are beyond reasonable expectations for providers to mitigate”.

It gave an example of not testing for whether a method could be side-stepped by having a parent or older sibling take the age check on a child’s behalf.

Another common example is using a VPN, a widely available web service that allows a user to funnel their internet traffic through other countries to access social media without the teen social media ban.

When France threatened to introduce age verification earlier this year and Aylo, the company that owns Pornhub and several other immensely popular websites, voluntarily blocked the country in protest, VPN services saw an immediate surge in demand.

The evaluation proposal plan also stressed that, even given its limited scope, it would not be able “test … all circumvention methods for all [Age Assurance] systems, due to the project’s timeline and available resources.”

Later, one member of the trial team would say that some circumvention testing was “much harder” to do in the trial testing and would require “policy response rather than technical measure”.

The limits on this circumvention testing was set by ACCS within the confines of the government’s tender, and confirmed by the government when they selected the group to carry out the trial.

The limited nature of this testing has been the biggest concern of the trial’s stakeholder advisory board, a group of more than 20 experts representing the spectrum of views from digital rights groups to anti-child exploitation organisations.

In every one of the minutes of three stakeholder advisory meetings that have been published, as well as a set of draft minutes obtained by Crikey, multiple members of the committee have questioned or registered concerns about how the trial is handling circumvention.

Rapid advances in AI and first-hand experience in children easily sidestepping methods were all raised as reasons to seriously consider further testing in the trial.

In a March meeting, one member of the advisory board, International Centre for Missing and Exploited Children Australia CEO Colm Gannon, said he was concerned that circumvention testing wasn’t a high priority.

“[Gannon] emphasised … that if the trial does not properly test for circumvention, the findings may lack credibility when applied to real-world implementation.”

The trial’s final testing for getting around the social media teen ban enforcement still isn’t known. A statement released today by the trial on its “preliminary” findings includes no information. The final report on the entire trial is scheduled to be given to the government at the end of July, who will choose what, if anything, will be released.

Even if all of that information is published, some of the circumvention testing details will be left intentionally opaque; ACCS CEO Tony Allen said the company wouldn’t disclose parts of the testing regime to avoid being exploited by bad actors.

Australia’s trial of the effectiveness of enforcing the teen social media ban has intentionally has been constructed in a way that means it won’t answer some of the key questions about its effectiveness.

But regardless of the trial’s scope, the teen social media ban will soon be put to the test. In just a few months, social media companies will be legally required to roll out these technologies to millions of Australians — and we will see whether 2025’s Tom will need even 30 minutes to get around the ban.

UNQUOTE

#AusPol #HahahahaLiebs #WhyIsLabor #NatsAreNuts #ClimateCrisis #WeAreTotallyFscked #GreensYeah #WomensRights #WomensRepresentation

VPNs and naughty parents: Teen social media trial isn’t testing some ways kids will get around the ban

Experts are concerned the trial may 'lack credibility' because it's not testing all the ways people could seek to circumvent the scheme.

Crikey

@MsDropbear42 @BinChicken @daedalus

"facial analysis tech tested by the trial"

Sure, Musk and Bezos needs a complete authorised list of all Australians including their faces.

Are there actually any brains at work?

#auspol

@petros @MsDropbear42 @BinChicken @daedalus It’s going to be really funny when people realise the way this will be implemented is that *everyone* will be required to give photo ID to Facebook, etc.
@MsDropbear42 @daedalus That's a great summary of the complexities of the problem, and a very effective multi-pronged response to what some government "experts" are claiming.
@BinChicken @daedalus Yeppy yep. Thing is, IMO, none of the problems were ever "mysterious", or left-field. Lots of we cynics were able to poke large holes in this foolishness right from the start... yet the pollies marched on, & the MSM slavishly followed, & afaict lots of non-techie parents felt hopeful... all sadly based on obvious bullshit. Really, peeps are just so silly, so often 🙄🤦‍♀️