Best wishes for a happy 2026! I hope you enjoyed a nice break!

You might notice the new name and look. AgeFriendly.AI is now Age Friendly Futures.

AI remains important to aged care's future, but it's one part of a broader picture. Tech adoption, policy, workforce development, cultural shifts, international innovation—these all shape the sector. Based on your feedback over the past couple of months, this newsletter now covers the full landscape.

If you're here for aged care strategy and innovation, you're in the right place. If you only wanted AI content, you'll still find it here, just not exclusively.

PS: I am currently in Xi’an, and this is a photo I snapped while strolling around early in the morning. A group of older people are doing exercises by following along on a video cast at a public exercise centre.

_{If you speak Mandarin, you will notice that some of the hanzi are a bit off, it’s because I took this photo from far away and then used AI to enhance it.}

China's Cyberspace Administration has released draft regulations for "human-like interactive AI services", i.e. AI designed to simulate personalities and form emotional connections with users. China is pushing for technology as a solution for many of the problems the aged care sector is facing, and this draft shows their effort to set the terms by strengthening safety and ethical requirements. You can access the translation here.

What the draft says about older users

The regulation names elderly companionship as an encouraged use case, but attaches specific obligations. Tech developers must prompt older users to nominate emergency contacts and alert those contacts if risks to life, health, or property emerge during use. Developers must also offer pathways to psychological support or emergency assistance and, more generally, products cannot be designed with goals of "replacing social interaction, controlling user psychology, or inducing addiction."

However, I feel the most interesting point about developing AI companions is that they should not imitate a user's relatives or other close relationships. This line rules out a design pattern common in companionship products, that is adopting a familial persona to build rapport with isolated older users.

Why it matters in Australia

China often moves first on technology regulation, and certainly early enough to surface the questions others will eventually face.

If you're designing or purchasing AI for older clients, this draft is worth reading as a checklist of issues regulators are starting to care about: dependency risk, emotional boundaries, and whether the product substitutes social connections.

We're currently planning a few exciting sessions. Keep an eye out for next week's newsletter where I'll share details on the first public sessions of 2026.

👉 Just in: I heard just yesterday that a provider is under investigation after a staff member used ChatGPT to write progress notes without de-identifying resident information. This hasn't hit the news yet, but I expect it will in the coming days.

This is exactly why one of the sessions we're developing addresses what staff are already doing with AI and the real choice ahead: implementing purpose-built systems with proper safeguards, or having staff continue using consumer tools that create compliance risks.

We'll cover how the right solutions deliver the support your staff need while meeting regulatory requirements. Stay tuned.

1 - Trading room and board for companionship 

Many older people live alone in large homes with unused bedrooms, while care and support workers face severe rental stress. A recent Washington Post opinion piece argues that current care and employment rules make it difficult to connect these two realities in a formal, affordable way, and proposes allowing housing to count as part of compensation. Offering a spare room in exchange for agreed support or companionship could reduce care costs, ease housing pressure for workers, and help older people remain at home.

2 - Predicting disease from sleep patterns

Stanford researchers have developed an AI model trained on 600,000 hours of sleep data that can predict over 130 health conditions—including Parkinson's, dementia and heart attacks with up to 89% accuracy by detecting when body signals fall out of sync during sleep.

3 - Paying in hours instead of dollars

Time banking—a mutual aid system where one hour of help equals one credit regardless of task—could address elder care challenges by making care more affordable, expanding the scope of support to include overlooked needs like companionship, and reducing social isolation. Researchers are exploring how to adapt the model for aged care, though challenges include sustaining participation and maintaining reciprocity when frail members can't easily give back.

4 - Copenhagen's universal design approach

The example of Copenhagen is discussed as an age-friendly city built through universal design choices (such as walkable streets, cycling infrastructure, public seating, and social spaces), rather than age-specific programs. These design choices keep people mobile, connected, and independent as they age, with research showing daily cyclists have 41% lower risk of premature death and walkable neighbourhoods reduce cardiovascular disease by 20-40%.

5 - Utah approves autonomous prescriptions refills by AI

In the US, the state of Utah is the first to allow an AI system to autonomously renew prescriptions for chronic conditions without doctor involvement, charging USD $4 per renewal for 190 eligible medications. 

Two roles from people I'd recommend working for:

Project Officer, Older Person's Care Hub – Bendigo Health's Continuing Care team (VIC) is recruiting for their new pilot across the Loddon Mallee region.

Head of ICT – SwanCare (WA) is looking for someone to lead strategic direction, development, and operations of their technology environment.

NotebookLM converts your data and research into infographics and presentation decks.

How it works step-by-step:

Load sources — Create a notebook and import files from Google Drive, paste links, or add text directly

Create infographics — Select "infographic" from the right panel, choose portrait or landscape, and adjust detail level and colours

Build presentations — Set deck length and language, then add instructions like "Compare performance metrics across three options"

Export — Download the deck or present directly from the platform

Every week, I'm looking into the changes shaping ageing and aged care, and sharing analysis worth your time. If this was useful, forward it to someone in the sector who'd appreciate it.

George Gouzounis