Read on for:
🌿 How Josh does his routine digital detox
📝 Jisoo’s exclusive interview with UQ Associate Professor Michael Noetel on Australia’s trust in AI survey
💻 Link to our Claude Cowork webinar
📫 Agency in the age of AI.
When I was a civilian policy adviser in Defence, I had the privilege of participating in a number of military exercises. Playing alongside my Army, Navy and Air Force colleagues, my role was to provide diplomatic and foreign policy advice to navigate a scenario as the 'blue team', pitted against the enemy on the 'red team'. These exercises were designed by a 'white cell' – a neutral team of planners writing the scenario, targeting our weaknesses and vulnerabilities to see how we would respond.
Sometimes, reading AI news feels exactly like this. A ‘white cell’ somewhere in the US or China is writing the scenario – and it can feel like we have no agency here in Australia. Yet we do. My exclusive interview with University of Queensland (UQ) Associate Professor Michael Noetel delves into this further.
Also a huge hello to our new subscribers. Please enjoy this welcome meme (made with the help of Gemini).
And to our existing subscribers, thank you so much for all your support and feedback so far. Please keep it coming as we continue to build out the ClearAI Community.
Now, let’s dive in to this fortnight’s drop.
Jisoo Kim
Co-Founder + Director
🇦🇺 AI News x Australia
Age Verification Rules: From 9 March, AI platforms are now required to comply with new rules preventing users under 18 from accessing harmful content. This follows Australia's world-first social media ban for teenagers and signals the government is extending its child safety framework into AI.
Comment: We are yet to see all 50 or so listed AI platforms fully comply but we’ll keep an eye on this now that the eSafety Commissioner can give out fines up to $49.5m (why not round it to $50m?) per breach.
Commonwealth Cousins Unite: Canada and Australia have strengthened the bilateral relationship, marked by Canadian Prime Minister Mark Carney’s visit Down Under. AI is a central pillar, with a Memorandum of Understanding (MoU) allowing our safety institutes to share expertise and work together.
Comment: Australia’s AI safety position is increasingly shaped by cooperation with likeminded democracies, more on this in The Feature.
Anthropic Fans Let Us Rejoice: Anthropic is opening an office in Sydney. We’ve been Claude fans for a couple of years now so it’s great to see it getting more mileage in Australia. As a nation, we currently rank 4th per capita for Claude usage.
🤠 The AI Round-Up
The Good
DermAI, a handheld tool developed by the University of Melbourne and promoted by CSIRO’s ON Accelerate program, is using AI to give GPs point-of-care specialist-level detection, particularly for diverse skin tones that often get missed by traditional algorithms. It’s a literal life-saver that puts high-end tech into the palms of rural doctors, making ‘equity in healthcare’ a real thing. Since 2 in 3 Australians face a skin cancer diagnosis by the age of 70, tools like DermAI are using intelligence in a way that benefits humanity and saves lives.
The Bad
Australia’s software juggernaut Atlassian has announced it is laying off roughly 1,600 staff (about 10% of its workforce). Atlassian has claimed that “AI use had changed the skills and roles the company needed, allowing a restructure to strengthen the company’s financial standing and self-fund further investment in AI and enterprise sales”.
Comment: We have looked at the employment stats at Atlassian over the last 6 years and wonder if overemployment bloat is the true cause of this cut. Sometimes top dogs like to use AI as an excuse for lay-offs but they often aren’t giving the full (or truest version of the) picture.
& The Ugly
CEO of OpenAI Sam Altman has suggested that ‘intelligence’ could soon be sold as as a utility where users pay for it on a meter. He said: “We see a future where intelligence is a utility, like electricity or water, and people will buy it from us on a meter. Fundamentally, our business and I think the business of every other model provider is going to look like selling tokens”.
Comment: Some dismiss this statement as a company justifying how it will make money going forward. We see it as a possibility given how AI usage hinges on energy, compute and data – all of which may likely to be much more expensive in the future.
🎟️ Events + 🎁 Goodies
🎟️ Claude Cowork Webinar
Tuesday 7 April, 10:30-11:15am AEST
Join us for a practical session on how Claude Cowork can automate your day-to-day file management, task handling, and business processes using nothing but basic language prompts. Oh, and it’s free!
🎟️ Black Swan Summit 2026
Josh and Jisoo will be speaking at separate AI roundtable discussions on Tuesday 24 March at the Summit in Perth. Shout out if you’ll be there. We’d love to say hi.
More info + tickets here.
🎁 Swapping to Claude? No wockas.
We really should have added this into the last drop but better late than never. A lot of people are telling us they want to switch to Claude but have given ChatGPT (or another LLM) so much context so they don’t want to start over. Luckily, the good eggs at Anthropic have made that easy. Check out the how-to here.
📝 The Feature
Continuing from my intro note…
So on my first military exercise, I was bogged down arguing the scenario was unfair, possibly unrealistic. My ever-patient and more experienced military colleagues told me it was best not to fight the scenario – the script handed to us by the white cell – which became a mantra I had to repeat to myself: "Don't fight the white." In doing so, I felt a loss of agency but I went along with it for the sake of the exercise.
In Australia, it can sometimes feel like there’s a ‘white cell’ somewhere in the US or China writing the scenario. We may sense we don’t have any agency. Yet we do. This isn’t a simulated exercise. This isn’t a game with a rigid script. We can change the outcome. We can vote with our feet, send strong signals to our leaders, say no to big tech, and prepare our institutions. We can fight the white.
I wanted to delve into this further, particularly after I caught up with Associate Professor Michael Noetel to discuss UQ’s latest Survey Assessing Risks from Artificial Intelligence 2025. Amongst key findings, the study found:
Australians want robust safety measures (89% would trust AI more if there were mandatory safety testing, independent audits, an AI Safety Institute [which the Government is currently setting up]), and
Demand for transparency and media coverage is strong (80-85% want more reporting on AI’s societal effects and on how government is managing AI regulation).
Nostalgic for a time when our computers looked like this? Same.
Image: Jisoo’s imagination x Gemini
I believe there are five main sentiments that drive the distrust underpinning the study’s findings:
We’ve been burned before.
Big tech promised social media would bring connection. Instead it delivered anxiety, addiction and polarisation. People aren’t being paranoid applying that same lens to AI.
It speaks like us and sounds correct. But it isn’t us and it can be wrong.
Natural language was made for humans so it’s disconcerting when a machine can speak it fluently, and also has the audacity to make mistakes so confidently with no ‘remorse’.
We can’t trust what we’re seeing anymore.
AI generated content is actively degrading the information environment we share. When hyperrealistic synthetic media ‘floods the zone’, healthy scepticism tips into not being able to believe anything. The shared reference points holding societies together come under assault.
We don’t trust other humans holding the tool.
AI doesn’t cut jobs. Employers do. AI doesn’t surveil citizens. Governments do. AI widens existing power asymmetries between the people making decisions and the people living with them.
We’re being told something is coming that we can’t imagine.
AGI. Superintelligence. Systems that exceed human capability across every domain. The people building it can’t agree on when or whether it’s safe or what it even means. Do you feel uneasy yet? But also, did anyone even ask for this?
Now, let’s jump into the interview here.
🕰️ The Analogue Edit
Digital detox – the Josh Phillips’ method
I’m an absolute sucker for my phone – Claude, ChatGPT, Gemini alongside Instagram and X. Somewhat shamefully, I’m more of a lurker than a contributor (who doesn’t love a good parenting meme). But this year my wife and I decided to try doing a digital detox once per month. The idea was to be present.
We (somehow) missed Jan and Feb, but the first Sunday of March we turned our phones off, put the laptops away and decided to go a full 24 hours without our devices.
Before I get to the experience, let me walk through the logistics. Like many households, we often begin on a Sunday: groceries, meal prep, coordinating the week by looking at the calendar and clearing my emails in the evening.
I use my phone for all of these. Woolworths app. ChatGPT for recipes. Calendar app. Superhuman for email.
So it required some prep.
We ordered our groceries on Saturday,
I used cookbooks and my memory to do our meal prep,
I did my calendar prep on Saturday night (including coordinating an early start with a colleague for Monday morning).
I also had to write my workout down on a notebook. Turns out I use AI for everything and my trusty Claw wasn’t going to be on hand when it came time to work out.
Sadly I had to accept the use of a streak freeze on my Duolingo.
So when Sunday rolled around, what happened?
Nothing, immediately. I got up, fed the children breakfast and got ready for the rest of the day.
It felt like time stretched, just a bit.
We were ready faster. I spent more time playing with the children. I read more than 100 pages of my book (Captain Corelli’s Mandolin if anyone’s wondering).
What stuck with me most is that I stopped caring about the many irrelevant pieces of noise that somehow force their way into my life on a daily basis. Life becomes simpler.
I have spoken with several others since who have also dared to spend time away from the screens. One doing it for the whole of his annual leave.
The verdict? They would all do it again. So will we. Actually, we have… every Sunday since so far.
If you’ve made it this far, thanks for joining us again. We’d love to know your thoughts – please hit us up. Otherwise, see you in the next drop.
Yours in humanity,
