26 June 2026

Welcome

Even though Australia has banned under-16s from social media, at least 4 in 5 of them are still there. The government now plans to build on this success with stronger restrictions, though they haven’t said what yet.

A British law firm claims to have won a case using an AI lawyer, while in Queensland students are protesting against a new AI data centre.

Everything issuing smoothly in this the best of all possible futures!

As this is the final The Weekly Cybers for the financial year, I’d like to thank the people who’ve thrown a few dollars in this direction. No one is paying me to do this (yet), so it’s much appreciated. Please feel free to join them, if you can.

Social media ban to face “further strengthening”?

Australia’s teen social media ban isn’t strong enough, says prime minister Anthony Albanese. “We’re working on that as a priority because this is something that other generations didn’t have to deal with, which is why it is complex,” he told parliament on Thursday.

Albanese referred to the ban as “one of this parliament's finest moments”, and said “we should be proud as a nation” that other countries are introducing similar bans.

“We can't allow the power that these companies, which are unaccountable, which get massive amounts of funding, of profit, and have extraordinary power — we need to make sure that Australians are in charge of this,” he said.

As we noted last week, for example, the UK is introducing a ban that’s been described as “Australia plus”.

Albanese’s comments follow the release of yet more research showing that the majority of under-16s — more than 85% in this case — were still using banned platforms, mainly through their own accounts.

Australia’s human rights commissioner Lorraine Finlay isn’t such a strong supporter of the ban, however.

“Young people are not simply passive recipients of protection; they are themselves rights-holders. Access to information, freedom of expression, and the ability to participate in democratic and social life increasingly take place in digital spaces. These are rights that matter not only to adults but also to children,” she wrote in The Mandarin this week.

“The social media ban risks doing real harm by shutting children out of the very spaces where they spend time with others, and begin to shape how they see themselves and the world... Our children deserve better than rushed responses to complex problems.”

Should we also ban older people from social media?

Meanwhile, research in Canada has shown that using social media is linked to poorer mental health in older adults, although as always “is linked to” doesn’t necessarily mean “causes”.

“Email use [was] positively linked with perceived mental health,” the researchers wrote in their full paper.

“No significant relationship was found between perceived mental health and the use of instant messaging apps, online voice or video calls, dating websites, or uploading self-created content.”

Also in the news this week

• From ASIO, we have the Director-General's Annual Threat Assessment 2026. There’s some intriguing tidbits about a network intrusion on a critical infrastructure operator, and some foreign spook activity in connection with AUKUS. “Using professional networking sites to recruit Australians is a low-cost and low-risk vector for foreign intelligence services," said ASIO boss Mike Burgess. I assume he means LinkedIn.

• Western Australia Police Force is deploying live facial recognition cameras to compare live images with a database of people with outstanding warrants, registered child sex offenders, and missing persons. That’s some 4,000 people in total. Needless to say, this isn’t going down well with privacy advocates.

• The federal government has introduced an AI review committee which consists solely of public servants and whose advice is non-binding.

• Sunshine Coast high school and university students have joined a protest against a joint Google-NEXTDC AI data centre under construction in Maroochydore.

• Is the Australian government in secret negotiations with AI companies over granting them copyright exemptions? A whistleblower claims they are, and independent senator David Pocock is supporting those claims. Industry and science minister Tim Ayres calls this “reckless speculation”, but he hasn’t actually denied the claims.

• AusAlert, the new national warning system, will be running a test on Monday 27 July, with an alert to be sent to every mobile device in Australia at 2pm AEST. Do not be alarmed.

END OF FINANCIAL YEAR SUPPORT?: The Weekly Cybers is currently unfunded. It’d be lovely if you threw a few dollars into the tip jar at stilgherrian.com/tip, or just forwarded it to others who might be interested. Thank you to those of you who’ve already done so.

Elsewhere

• Last week Elon Musk was a trillionaire. This week Elon Musk is not a trillionaire. Meh, it’s all pretend money anyway.

• The major AI platforms are facing their big challenge: convincing companies to pay what all this coming power is actually costing them. Many tech workers were told to maximise their use of AI. Now that the price of so-called “compute tokens” is soaring they’re being told to cut back (New York Times gift link).

• The Five Eyes nations’ signals intelligence agencies have issued a joint statement warning that AI models capable of devastating cyber attacks on governments and business are mere months away. “We must act swiftly to remain ahead,” they write.

• China’s LineShine has overtaken US-based El Capitan as the world’s most powerful supercomputer, according to the TOP500 leader board. Getting the top spot is perhaps more about bragging rights than anything else, but it will certainly provide more grist for the narrative mills to repeat that the West is falling behind.

• The US Federal Communications Commission (FCC) is to create tougher rules for submarine cables to make it harder for Chinese companies to provide equipment and to fast-track approvals for American firms.

• China Digital Times has an update on the intolerance of the Chinese internet, and the narrowing of online discourse to politically safe topics.

• It may be a world first? AI-powered law firm Garfield AI claims to have won a case using AI — although an actual human, a junior barrister named Dominic Li, had to front up to Wandsworth County Court in England.

• Here’s some interesting suggestions for how to talk about AI without anthropomorphising it. Some of them are as basic as replacing “prompt” with “text input”, and “answer” with “output”.

• In related news, it turns out you could build a ChatGPT-like large language model (LLM) using virtual goats in Age of Empires II. The aim of the researcher’s paper is to disavow us of the idea that LLMs can exhibit consciousness or even “thinking”.

• A software engineer in North Carolina has claimed a religious exemption from using AI at work. So far it’s just a claim, but now the employer has to show that it’s unsupportable. Definitely one to watch.

• “Companies are adopting AI before they're ready to do so and are paying the price”, reports The Register. OK, it’s from a vendor’s research, so take the exact numbers with a very big grain of salt, but they reckon that 93% of organisations that have taken the plunge have reported “infrastructure incidents” attributable to AI.

• Tech Policy Press asks: Who really controls your digital likeness in the age of AI wearables?. The answer is: Not you.

Inquiries of note

• When I mentioned last week that the Australian Signals Directorate (ASD) had issued a new version of their Information Security Manual (ISM), I forgot to mention that they’re also running a consultation to morph their award-winning Essential Eight cybersecurity framework into a broader “Essentials” series. “Consultation on Essentials for enterprise IT is now open via the ASD Cyber Security Partnership Program portal and will run until 12 July 2026,” the ASD writes. Additional chapters will follow.

What’s next?

Parliament continues this coming Monday 29 June to Thursday 2 July, after which it’s on its usual long winter break until 11 August.

DOES SOMETHING IN THE EMAIL LOOK WRONG? Let me know. If there’s ever a factual error, editing mistake, or confusing typo, it’ll be corrected in the web archives.