CHINA INSIGHTS | Part II
8 December 2025
Robots in the Cupboard: A Warning about Australia's Regulatory Caution
A young girl in Shenzhen runs after a small robodog moving across a public plaza, December 2025.
Over the past week in Shenzhen, I've spoken with numerous robotics manufacturers about their products for aged care. I've trialled exoskeletons, tried a stair-climbing wheelchair, was led by a guide robodog. But this is far from the most impressive part of my stay. Because Shenzhen lives in the future.
I've been driven by driverless taxis (the first time feels creepy – I was genuinely stressed), watched the steering wheel turn itself through busy intersections, and arrived at my destinations only a few minutes late each time, as my car kept giving way to human drivers. The fare was half what I'd have paid for a regular taxi. At Talent Park, I watched drones descending with food delivery orders for people picnicking. This afternoon, a robot made me a cappuccino with latte art, my face printed on it.
I'm the type of person who talks to baristas and taxi drivers. I enjoy those brief connections—the barista who remembers my order, the driver who recommends a restaurant. I'll miss that if robotics become ubiquitous in my lifetime. And yet, watching that robot make my coffee, I couldn't deny the appeal. Lower costs. Consistent service. No awkward interactions for those who wish to avoid them. Just the thing you ordered, appearing when you need it.
That tension—between appreciating what technology can do and mourning what it replaces—sits at the heart of robotics in aged care. We're not facing a simple choice between human warmth and technological efficiency. We're facing decisions about which tasks genuinely require human judgement and which don't.
In Australia's case, the real problem isn't that we lack the will—it's that regulatory caution prevents us from adopting technologies already proven elsewhere. We've acknowledged this reality for medical devices by (sort of) accepting overseas approvals, but there’s a lot more that can be done. Meanwhile, Japan has taught us an expensive lesson about what not to do, and China is showing us what strategic ambition looks like.
When marketing meets physics
Exoskeletons designed to help aged care staff move through facilities and assist with manual handling could reduce workplace injuries and physical strain.
The exoskeleton marketing video I was shown featured fit models walking and running with fluid, natural grace. The actual experience, however, felt like learning to walk again, and I even had trouble walking backwards. My movements weren't quite right.
The stair-climbing wheelchair, by contrast, simply worked. No flashy marketing needed. It solved a specific problem: giving people access to spaces they couldn't previously reach. This is the model for good robotics adoption: identify the problem, deploy technology that solves it, move on.
I've been driven by driverless taxis (the first time feels creepy – I was genuinely stressed), watched the steering wheel turn itself through busy intersections, and arrived at my destinations only a few minutes late each time, as my car kept giving way to human drivers. The fare was half what I'd have paid for a regular taxi. At Talent Park, I watched drones descending with food delivery orders for people picnicking. This afternoon, a robot made me a cappuccino with latte art, my face printed on it.
I'm the type of person who talks to baristas and taxi drivers. I enjoy those brief connections—the barista who remembers my order, the driver who recommends a restaurant. I'll miss that if robotics become ubiquitous in my lifetime. And yet, watching that robot make my coffee, I couldn't deny the appeal. Lower costs. Consistent service. No awkward interactions for those who wish to avoid them. Just the thing you ordered, appearing when you need it.
That tension—between appreciating what technology can do and mourning what it replaces—sits at the heart of robotics in aged care. We're not facing a simple choice between human warmth and technological efficiency. We're facing decisions about which tasks genuinely require human judgement and which don't.
In Australia's case, the real problem isn't that we lack the will—it's that regulatory caution prevents us from adopting technologies already proven elsewhere. We've acknowledged this reality for medical devices by (sort of) accepting overseas approvals, but there’s a lot more that can be done. Meanwhile, Japan has taught us an expensive lesson about what not to do, and China is showing us what strategic ambition looks like.
When marketing meets physics
Exoskeletons designed to help aged care staff move through facilities and assist with manual handling could reduce workplace injuries and physical strain.
The exoskeleton marketing video I was shown featured fit models walking and running with fluid, natural grace. The actual experience, however, felt like learning to walk again, and I even had trouble walking backwards. My movements weren't quite right.
The stair-climbing wheelchair, by contrast, simply worked. No flashy marketing needed. It solved a specific problem: giving people access to spaces they couldn't previously reach. This is the model for good robotics adoption: identify the problem, deploy technology that solves it, move on.
The robotic coffee machine that made me a coffee and printed my portrait onto the foam.
Japan's $300 million cautionary tale
Japan seemed perfectly positioned to lead in care robotics. Over two decades of development, more than $300 million in government funding by 2018, extensive expertise in industrial robotics, and a acute demographic crisis.
Yet in 2019, only 10% of Japanese elder care institutions reported using any care robot. Among home care providers, just 2% had experience with them. When robots were purchased, they often ended up locked in cupboards after brief trials.
Ethnographic research by James Wright at a Japanese nursing home reveals why. Staff stopped using a lifting robot within days—it was cumbersome to wheel between rooms, cutting into their time with residents. A robotic seal for companionship created its own problems: one resident became so attached she refused to eat or sleep without it, forcing staff to constantly monitor the interaction. A humanoid robot meant to run recreation sessions required staff to stand beside it, copying its movements to get residents to participate.
The pattern was consistent: robots didn't save labour, they just changed its nature. Care workers spent less time building relationships with residents and more time managing, maintaining, cleaning, booting up, and monitoring machines.
This is the disaster Australia must avoid. Not the failure of technology itself, but the failure to ask the right questions before procurement: Will staff actually use this? Does it create more work than it saves? Does it enhance the relational aspects of care?
China's strategic response
Here in Shenzhen, the future feels closer than anywhere else I've visited. Hundreds of companies manufacture humanoid robots in China, with more entering the market constantly. Last week, China's National Development and Reform Commission issued a warning about "highly similar" models overwhelming the market and squeezing out space for meaningful research and development.
This may sound cautionary, but it's strategic planning. The same agency naming humanoid robotics as one of six key economic growth drivers through 2030 is also warning against unfocused investment. They're accelerating efforts to build market mechanisms for entry and exit, whilst further supporting R&D for core technologies.
This is industrial policy done properly: encouraging innovation whilst preventing wasteful duplication that leads to one hundred nearly identical products competing for narrow applications. Australia should pay attention to this approach—not the scale of investment (we can't match that), but the combination of ambition and discipline.
China is testing robotics at scale precisely because they face the same workforce crisis we do. They're throwing capital at the problem whilst maintaining strategic oversight. Australia, meanwhile, debates processes whilst falling further behind on practical technology adoption.
Japan seemed perfectly positioned to lead in care robotics. Over two decades of development, more than $300 million in government funding by 2018, extensive expertise in industrial robotics, and a acute demographic crisis.
Yet in 2019, only 10% of Japanese elder care institutions reported using any care robot. Among home care providers, just 2% had experience with them. When robots were purchased, they often ended up locked in cupboards after brief trials.
Ethnographic research by James Wright at a Japanese nursing home reveals why. Staff stopped using a lifting robot within days—it was cumbersome to wheel between rooms, cutting into their time with residents. A robotic seal for companionship created its own problems: one resident became so attached she refused to eat or sleep without it, forcing staff to constantly monitor the interaction. A humanoid robot meant to run recreation sessions required staff to stand beside it, copying its movements to get residents to participate.
The pattern was consistent: robots didn't save labour, they just changed its nature. Care workers spent less time building relationships with residents and more time managing, maintaining, cleaning, booting up, and monitoring machines.
This is the disaster Australia must avoid. Not the failure of technology itself, but the failure to ask the right questions before procurement: Will staff actually use this? Does it create more work than it saves? Does it enhance the relational aspects of care?
China's strategic response
Here in Shenzhen, the future feels closer than anywhere else I've visited. Hundreds of companies manufacture humanoid robots in China, with more entering the market constantly. Last week, China's National Development and Reform Commission issued a warning about "highly similar" models overwhelming the market and squeezing out space for meaningful research and development.
This may sound cautionary, but it's strategic planning. The same agency naming humanoid robotics as one of six key economic growth drivers through 2030 is also warning against unfocused investment. They're accelerating efforts to build market mechanisms for entry and exit, whilst further supporting R&D for core technologies.
This is industrial policy done properly: encouraging innovation whilst preventing wasteful duplication that leads to one hundred nearly identical products competing for narrow applications. Australia should pay attention to this approach—not the scale of investment (we can't match that), but the combination of ambition and discipline.
China is testing robotics at scale precisely because they face the same workforce crisis we do. They're throwing capital at the problem whilst maintaining strategic oversight. Australia, meanwhile, debates processes whilst falling further behind on practical technology adoption.
A drone is lowering a food delivery box toward a landing pad at Shenzen's Talent Park, December 2025.
The regulatory barrier—and the obvious solution
Australia's super slow reaction is frustrating because we've already acknowledged that duplicating regulatory assessments done by comparable jurisdictions is wasteful. The Therapeutic Goods Administration accepts CE marking from the European Union and FDA approvals from the United States. This recognition exists because Australia's market is too small to justify comprehensive independent assessment of every device. And yet, a stair-climbing wheelchair with CE marking shouldn't wait six months for TGA listing.
The solution is straightforward: expand mutual recognition beyond merely accepting certifications as “supportive evidence” to include automatic TGA listing for devices already approved for aged care applications in comparable jurisdictions. If a robotic device has been successfully deployed in aged care facilities across the EU or North America for two years with documented safety records and staff satisfaction, Australian providers shouldn't face additional regulatory delays.
We need a fast-track category specifically for assistive technologies and care robotics where overseas classification and approval are already established.
A final thought
I'll miss conversations with taxi drivers as autonomous vehicles become standard. But here's the difference: coffee and transport are optional conveniences where we can prioritise human interaction if we value it. Aged care isn't optional. And aged care workers shouldn't be delivering linens when they could be having chats.
The choice is between thoughtful, disciplined adoption of proven solutions and continued inaction disguised as careful consideration. Japan chose to spend first and ask questions later—they got robots in cupboards. Australia is choosing to ask endless questions and spend never.
There's a third option: learn from both. Test rigorously. Adopt proven logistics solutions quickly. Streamline regulatory recognition of devices already proven elsewhere.
The robots are here. The question is whether we're ready for them.
Australia's super slow reaction is frustrating because we've already acknowledged that duplicating regulatory assessments done by comparable jurisdictions is wasteful. The Therapeutic Goods Administration accepts CE marking from the European Union and FDA approvals from the United States. This recognition exists because Australia's market is too small to justify comprehensive independent assessment of every device. And yet, a stair-climbing wheelchair with CE marking shouldn't wait six months for TGA listing.
The solution is straightforward: expand mutual recognition beyond merely accepting certifications as “supportive evidence” to include automatic TGA listing for devices already approved for aged care applications in comparable jurisdictions. If a robotic device has been successfully deployed in aged care facilities across the EU or North America for two years with documented safety records and staff satisfaction, Australian providers shouldn't face additional regulatory delays.
We need a fast-track category specifically for assistive technologies and care robotics where overseas classification and approval are already established.
A final thought
I'll miss conversations with taxi drivers as autonomous vehicles become standard. But here's the difference: coffee and transport are optional conveniences where we can prioritise human interaction if we value it. Aged care isn't optional. And aged care workers shouldn't be delivering linens when they could be having chats.
The choice is between thoughtful, disciplined adoption of proven solutions and continued inaction disguised as careful consideration. Japan chose to spend first and ask questions later—they got robots in cupboards. Australia is choosing to ask endless questions and spend never.
There's a third option: learn from both. Test rigorously. Adopt proven logistics solutions quickly. Streamline regulatory recognition of devices already proven elsewhere.
The robots are here. The question is whether we're ready for them.