12 June 2026

Welcome

New research shows a clear link between heavier teen social media use and an increase in mental health problems, though we don’t know how exactly.

AI continues to dominate the news, of course. OpenAI has joined Anthropic and SpaceX (yes, an AI company too) in preparing for a share float, while Google is issuing more shares.

The Australian government seems to believe in the boom, as is the entire world it seems. I’m sure it will end well.

Teen social media use linked to depression

A decade-long study led by Murdoch Children’s Research Institute (MCRI) has confirmed a link between heavier teen social media use and symptoms of depression and anxiety — although we can’t be sure which way the causality flows.

The study tracked about 1,200 Melbourne kids from 2012 and 2013, and compared the mental health of those who used social media for less than an hour a day with those who used it for two hours or more.

“More than two hours of daily social media use was associated with an increased risk of high depressive symptoms and poor well-being one year later. Estimated risks were greater in early adolescence (12–13 years) compared to risks in later ages,” the researchers wrote.

Broadly speaking, for every 100 teens using social media for more than two hours, somewhere between 6 and 11 more of them were showing various symptoms when compared with those using it for less than an hour. The difference was greater for girls than for boys.

While this may not sound a lot, it’s worth remembering that until the teen social media ban came into force — well after this data was collected — typical social media use was around two and a half hours per day.

“The effects were modest at the individual level, but because social media use is so common among adolescents, even modest increases in risk could translate into potentially meaningful impacts at the population level,” Dr Nandita Vijayakumar told a media briefing on Wednesday.

“Importantly, levels of social media use among adolescents have increased over time, so over two hours of use is widely prevalent nowadays, meaning that the population impact could be even greater.”

The increase in the relative strength of algorithmic engagement has also increased, the researchers said, meaning that the figures could now be higher.

Meanwhile The Conversation has published five tips to make social media less addictive and, more broadly, six steps to safer social media.

Datacentre boom too big to ignore: Charlton

Australia’s datacentres boom may be bigger than the resources boom, and despite the concerns over energy and water consumption, we can’t ignore the economic wave, according to Andrew Charlton, Assistant Minister for Science, Technology, and the Digital Economy.

“Comparable surges, such as railways, electricity, and telecommunications rolled out over generations. This one is similar in size, but is being compressed into a single decade,” he said in a speech to the Sydney Institute this week.

“In the nineteenth century, economic power rested on coal. In the twentieth, it rested on oil. In the twenty-first, it will rest on computation; and data centres are where computation is made. AI is the story. Data centres are simply where the AI happens.”

Charlton pointed to an obvious failure of the resources boom, when we had abundance, “but we didn’t lock in our advantage early enough”.

“We built for export before securing supply at home,” he said.

This problem was explained recently by If You’re Listening.

The government has been accused of an AI policy retreat, however. Labor’s former minister Ed Husic says the government “blinked” on regulation, dumping his plans for “mandatory guardians” on high-risk AI applications.

The topic was discussed on ABC TV’s Four Corners this week, The AI Race.

Also in the news this week

• The Department of Home Affairs wants to give telcos and cloud services providers the power to block cyber threats further upstream, which may require “enhanced legislative and policy levers” — which is modern policy-speak for “new laws”. The proposal is part of the department’s Horizon 2 Action Plan (PDF), itself part of the 2023–2023 Australian Cyber Security Strategy.

• Employment services providers have been submitting their AI plans to the Department of Employment and Workplace Relations (DEWR), but there’s not much action yet. “About six months after most of those plans were submitted, none have been approved,” reports The Mandarin.

• Dozens of government bodies have failed to police their AI use. More than half the agencies check by the Digital Transformation Agency (DTA) had failed to meet their first deadline for transparency.

• The Australian National Audit Office (ANAO) has published a report on the Management of Cyber Security in the Department of Parliamentary Services (DPS). It reveals that the federal parliament computer network will receive its “most significant” upgrade since being built, because the current network is “outdated, unsegmented, and increasingly vulnerable to interference and disruptive cyber events”.

• Ever-reliable Services Australia has sent pensioner concession cards to the wrong people.

• Optus has apparently breached 51,000 customers’ privacy by publishing their unlisted phone numbers in something called the White Pages. The problems go back years. We don’t yet know how big a fine they’ll cop.

• Chatbots keep telling stories about a lighthouse keeper named Elias Thorne. We might know why, reports 404 Media.

PLEASE SUPPORT THIS NEWSLETTER: 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

• ChatGTP creator OpenAI has filed for an IPO, joining Anthropic and SpaceX. While the documents aren't public yet, OpenAI was last valued at US$852 billion. The company has also written a blog post about its vision for an AI future.

• As I mentioned up the top, Google is issuing more shares too, in part to build more datacentres.

• Anthropic has released a new AI model, Claude Fable 5, which is supposedly a version of Claude Mythos 5 that’s safe for us plebs to use. I am chuckling quietly.

• Around 30 billion “environmental scan” images snarfed from the phones of Pokémon Go players have been used to build a navigation model for military drones.

• A German court has ruled that Google is liable for false statements in its AI Overviews.

• Google and Meta have been denied new a trial in that Californian court case about youth social media addiction.

• The big tech companies are trying to convince us that opposition to data centres is a Chinese influence operation. But as the Australian Strategic Policy Institute (PSI) notes, “so far, this research all comes from organisations that are openly aligned with the AI industry or the aggressively pro-AI Trump administration, not from organisations that specialise in China or foreign influence”.

• Apple announced AI upgrades to Siri, expected later this year, but they may not be available in Europe (New York Times gift link).

LATEST PODCAST: If you’re reading this newsletter — and you are — I suspect you’re the kind of person who’d also be into the argument over whether Pluto is or isn’t a planet. We discuss that in The 9pm Rockets are Always Blowing Up with Dr Alice Gorman and Rami Mandow. Look for The 9pm Edict in your podcast app. We also talk about new ways to make black holes, glitches, drones, the explosion of Blue Origin’s New Glenn launch vehicle, and of course the SpaceX share float.

Inquiries of note

Nothing new this week.

What’s next?

Parliament is now on a break until Monday 22 June, which is 10 days away.

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.