When my sister and I were kids, my parents had a tiny beach shack. Two doors down lived an older fisherman, bearded, burly, and weathered. He cured his own fish, significant quantities of it, and had it throughout the day with a few glasses of ouzo. One summer my sister and I found his stash and kept raiding it daily until he caught us red-handed. I remember him dearly, and I remember the bottle of ouzo was always nearby.

The picture is familiar to many of us. An uncle, a grandfather, or a neighbour, having an afternoon glass, then an evening glass, the bottle always on the bench. For most of us at the time, it read as part of the landscape, however the data tells a different story. People in their 60s now drink at risky levels more often than the general adult population. One in nine Australians aged 70 and over drinks every day. 60% of alcohol-induced deaths in 2024 were people aged 55 and over. Alcohol use disorders are linked to a roughly threefold higher risk of dementia, and alcohol is one of the few dementia risk factors anyone can do something about. Most of the people inside these numbers are ordinary Australians whose drinking habits were set forty or fifty years ago.

The Australian Communications and Media Authority is currently reviewing the rules on alcohol advertising on television, with submissions open until Thursday this week. Recent polling has 75% of Australians wanting fewer alcohol ads on TV. Worth a submission if you have a view.

👉 A note for new subscribers from the Invox Support at Home Conference and the Anglicare Sydney session this week — welcome. Since you’re new, here are some of the most popular recent writeups: The three tiers of AI in aged care | What I observed in China on ageing, aged care, and technology | Cultural intelligence with Robert Bean | Capital asymmetry with Prof Rick Watson

Alice Wang is the Clinical Services Manager at AVWA in Melbourne. A month ago she had a problem that many of us have faced. A policy change had to reach more than 300 frontline staff across multiple sites. The standard approach had been an email with the policy pack attached; in this case, 15 (!) documents including FAQs. Aged care managers know what tends to happen with those emails. Most go unread, and even for those at a desk, dense policy material is the hardest format to absorb a change from.

Face-to-face sessions across multiple sites would have cost time and budget that most providers run thin on.

So Alice tried something different. She loaded the 15 documents into AI and used it to produce an eight-minute training and launch video. She gave it the framework and structure: context first, then what was changing, then specifics, then the active date. She was kind enough to share the output with me and I was impressed. Visual, sequenced, easily watchable on a phone, and all Alice with staff was one link. The consistent message reached every staff member regardless of site or shift.

Alice had come to a session I ran at Moonee Valley Connect a few weeks earlier. The session covered several tools, but the specific tool is secondary to the principle: AI works best when you stop using it to do faster versions of what you already do, and start using it to do the thing you actually need done. Alice took that and applied it to a problem her organisation had been working around since forever.

NotebookLM Plus is what Alice happened to use. The tool is the smaller part of the story. The bigger part is one person identifying a clear problem in her own work and giving herself a fortnight to try something on it.

With thanks to Alice for sharing this.

Assist by i-P (part of the AgedCareAI suite) is the first web AI extension built for Aged Care and NDIS organisations. It's now on the Chrome Web Store.

Most aged care AI sits in a separate tab or tool, and only sees clinical data. Assist works differently. Install it once and it lives in the side panel of every web-based system your team already uses — clinical, rostering, finance, HR, nurse call, maintenance, complaints, risk, compliance, government data connections, and whatever else you run on.

It sees and understands the form or page in front of you, pre-fills it from your other systems, and answers questions in plain language using your own templates, policies and procedures.

Assist is trained on your organisation, for your organisation. It runs in a secure environment with no link to public AI tools. Your data is never shared, never used to train other models, and never leaves your control.

Tuesday 19 May | 10.30am AEST | Online | Free

A live Q&A with Elyssia Clark (Benetas), Justina Gardiner (LightSpeed Consulting), and Hanaa' Grave (culturalQ) on who's actually making aged care decisions in 2026, how cultural diversity shapes the consumer base, and where AI fits.

The aged care consumer in 2026 is more than one profile. Decisions move through families with different priorities and unequal dynamics. Cultural diversity is shifting too, and and consumers are arriving more informed — 66% of families say online reviews shape their decision. The combination is what's new, and the pace it's changing at.

6–7 May | Royal International Convention Centre, Brisbane

I'll be part of the closing session — a fishbowl-style panel called "Aged Care 2076: Wild Predictions and Bold Ideas" with Sanka Amadoru (Geriatrician, Aria Health) and Dan Aitchison (CEO, Palm Lake Care). We're throwing out the rulebook and jumping 50 years into the future to imagine what aged care in Australia could look like in 2076. Expect provocation, energy, and a few ideas that will either inspire you or make you nervous.

A new study by three universities followed more than 2,100 dementia-free Australians, middle-aged and older. For every 10% increase in ultra-processed food (UPF) intake — about a packet of chips a day — participants scored measurably lower on tests of visual attention and processing speed. The detail worth noting: the effect held even for people otherwise eating a healthy Mediterranean diet. The processing itself appears to matter as a mechanism in its own right. Average UPF intake among participants was 41% of daily energy, almost at the national average of 42%.

Anny Junek has won the Purim costume contest at her Israeli retirement home three times. She's 94, lost her parents in Auschwitz, survived Bergen-Belsen, raised a family in Mexico, and has a passion for costumes. The short documentary by Tamir Elterman was filmed in 2019, but I came across it last week and thought it was worth sharing.

Amplitude published their Australian survey data this month. Only 4% of workers aged 55–64 say they trust AI recommendations over their own judgement, compared with 31% of those aged 18–24. Daily AI use shows the same split — 39% versus 20%. The report draws a straightforward implication: older workers tend to sit in leadership roles and remain the most sceptical, while younger workers use AI daily but lack the mandate to shape how it gets adopted. Aged care has a similar age structure across leadership and frontline roles, so this gap is probably shaping a lot of the adoption decisions being made right now.

This is why I keep encouraging managers to ask their staff a simple question: what are you already using? That conversation is where adoption actually starts.

A week after I presented you a workflow for building diagrams in Gamma, ChatGPT released their new image model (access it in your account) that takes the image game to the next level.

I tried it with a much simpler prompt that I'd used in Gamma: "create an infographic about falls prevention for older people." Just one prompt, and the output is a usable poster. (see below)

What's new: the model renders text inside images reliably (this was the main limitation before), handles dense layouts, and works inside the standard ChatGPT interface. This means quick first-draft diagrams, workflows, posters, training slides, education flyers you can refine in conversation. For finished work that needs polish, a designer is still where you go, but for an internal poster, it'll save you ten minutes in Canva.

Remember to iterate by asking it to change one thing at a time. "Make the heading smaller." "Use warmer colours."

The better "thinking" mode is on paid plans, but the standard mode, which is what I used, is on the free ChatGPT plan and it’s impressive still.

Each week, I review developments in ageing and aged care and what they mean in practice. If this was useful, forward it to someone in the sector who'd appreciate it.

George Gouzounis