As South Africa’s elderly population grows, a new generation of AI companions and robots is moving into the home, raising the question: can a machine help care for our ageing parents when we cannot?
There is a particular kind of guilt and intense worry that comes with having an ageing parent living alone. Who hasn't got into a panic when a call remains unanswered. Worries hit deep: Are they okay? Have they eaten? Did they fall? Are they lonely? Did they take their medication?
For millions of adult children, it is an increasingly familiar anxiety.
And it is an anxiety South Africa will increasingly have to confront.
South Africa’s ageing population
Statistics South Africa says the number of people aged 60 and older rose from about 3.6 million in 2002 to 6.6 million in 2025. Older people now make up approximately 10.5% of the country's population. Stats SA says this proportion is projected to reach 16% by 2050.
The Western Cape is particularly significant. Census 2022 found that about 11% of the province's population was aged 60 and older, one of the highest proportions in the country.
At the same time, families are increasingly dispersed.
Children may work in another province, live overseas or spend most of their day at work.
An elderly parent may still be perfectly capable of living independently but need someone to remind them about a tablet, an appointment, a meal or simply to ask how their day is going.
That is precisely the territory AI companions are moving into. The technology is not necessarily trying to keep an elderly person alive but rather help them stay independent for longer. And to me that sounds like a really good idea.
Enter the technology industry, not with another smartphone app or another complicated medical device, but a voice in the room.
What these companions can do
The latest generation of artificial intelligence is being built to talk to older people, remind them about medication, help structure their days, provide entertainment and conversation and, in some cases, quietly give their families an indication that everything appears to be normal.
One example is Ato, a voice-first AI companion designed specifically for older adults.
News about it landed as a newsletter in my inbox late last night and I went to research it.
The idea is simple: plug it in and talk. There is no smartphone screen to navigate, no complicated menu system and no password to remember. Perfect.
According to the company, Ato can hold conversations, answer questions, provide entertainment and deliver reminders for medication, appointments and everyday routines. Its companion app allows family members to send messages, schedule reminders and monitor activity.
That simplicity is important for an older person who has spent a lifetime communicating by telephone rather than touchscreen, or who struggles with the fine motor skills required to operate a smartphone, the ability to simply say what they need can make technology accessible in a way an app never will. It also is part of push for voice-enabled tech.
And Ato is not alone.
ElliQ, another AI companion designed for older adults, was also part of the early morning newsletter.
The company says it combines conversation and entertainment with medication reminders, health and pain tracking, exercise prompts and communication with family members. Its caregiver app allows relatives to send messages and photos, make video calls, set reminders and receive alerts when a user reports problems such as pain or changes in mood.
The companies behind these products say they are not selling them as replacements for doctors, nurses or family members.
But there is a catch for South African families.
ElliQ, one of the best-known AI companions for older people, is currently available only in English in the US and ships and provides support exclusively there. It also requires both electricity and Wi-Fi to operate. Its privacy policy says the device uses microphones and sensors, but does not record conversations in the background when idle, while personal data is not monetised.
That illustrates one of the problems with importing this technology into South Africa: affordability and availability are only part of the equation. Connectivity, electricity, language and data privacy matter too.
The proposition is considerably more modest and potentially more important. They are trying to fill the gaps between human visits.
In simple terms, the emerging market looks something like this:
Voice companion Conversation, reminders, alerts and companionship.
Physical robot Emerging technology, expensive and with limited availability.
Human care Judgement, physical assistance, care and emotional support.
The care crisis behind the technology
The biggest driver of the industry is not gadgets but the maths.
The world's population is ageing while the number of people available to provide hands-on care is under pressure.
Japan is one of the clearest examples I found in my research. Its rapidly ageing population has made care-worker shortages a national problem and turned the country into one of the world's most important testing grounds for care robotics.
But something unexpected has emerged. New research from Stanford, the University of Tokyo and the University of Notre Dame examined the effect of robots in Japanese nursing homes.
The study found that nursing homes adopting robots subsequently recorded a 28% increase in care-worker employment, a 39% increase in nurse employment and about 26% higher total employment. The increase occurred among non-regular employees, such as staff working under more flexible or part-time arrangements. Robot adoption was also associated with fewer staff-retention problems.
The researchers say the findings suggest robots can complement rather than simply replace caregivers. The results are from the Japanese care sector, however, and should not be taken to mean robots will create jobs universally.
The researchers argue that robots can take over particular tasks while allowing human carers to spend more time on work requiring what they describe as the “human touch”.
That is potentially one of the most important findings in the emerging debate over AI and employment.
My first instinct was to ask: Will the robot replace the caregiver? But the Japanese evidence from this study suggests not.
What, instead, if the robot makes it possible for the caregiver to spend more time actually caring?
A machine can remind someone to take a tablet but cannot replace the comfort of holding their hand. A machine can monitor a routine but it cannot replace the warmth of a conversation with a daughter who has driven across town to visit. A machine can help with repetitive work but it cannot replace human affection.
At least, not yet.
China is betting heavily on the machines
China has recognised the same demographic pressure and is moving aggressively.
In 2025, China's Ministry of Industry and Information Technology and Ministry of Civil Affairs launched a national pilot programme for intelligent elderly-care robots, running from 2025 to 2027.
The programme covers care in homes, communities and institutions, including support for people with disabilities or cognitive impairment, emotional companionship, health promotion, smart-home functions and everyday assistance.
For home-based trials, the programme requires participating projects to test robots in at least 200 households, with at least 200 robot units deployed, over a validation period of at least six months. It is a pilot programme designed to test and improve the technology, rather than evidence that robots are already widely deployed in Chinese homes.
It is a striking indication of where the industry believes the next major robotics market lies: not in factories but in living rooms, bedrooms, kitchens and eventually, perhaps, the bathrooms and bedrooms where much of the hardest physical caregiving takes place.
From talking boxes to humanoids
At the moment, the most practical end of the market looks almost mundane.
A device sits on a table, you talk to it and it talks back.
But at the other end of the industry, companies are trying to build machines that can physically move around a home.
1X's NEO, for example, is being marketed as a general-purpose home humanoid. The company currently lists an Early Access ownership price of $20,000, with a $200 refundable deposit and US deliveries starting in 2026. It also offers a $499-a-month subscription.
I checked and there is currently no South African delivery option listed by 1X: its order page specifies US deliveries starting in 2026. At the company's listed Early Access price of $20,000, the robot would cost roughly R330,000 at an exchange rate of R16.50 to the dollar, before shipping, taxes and any local costs. So cost is going to be a thing.
The humanoids are a very different proposition from an AI speaker.
A voice companion can remind an elderly person to take medication, while a physical robot could potentially fetch medication in the future. A voice companion can ask whether someone has eaten, while a physical robot could eventually prepare or bring food and a voice companion can call a family member, while a humanoid could potentially help someone stand up, carry objects, open doors or perform other physical tasks as the technology develops. It's what my sister and I have dreamt about.
These are potential future uses of humanoid robots, rather than established functions currently listed by 1X for NEO. NEO is being developed as a general-purpose home robot, but its capabilities should not be confused with the broader possibilities being imagined for elder care - which I will watch closely.
That is where the technology becomes both exciting and deeply uncomfortable.
The privacy and human-cost questions
There is a point at which an AI companion stops being a gadget and starts becoming part of a person's social world.
And that raises difficult questions. If an elderly woman spends hours every day talking to an AI because she is lonely, is that a technological success or evidence of a society that has failed to provide her with enough human company?
If a family uses a robot to check on a parent remotely, does that provide valuable independence, or does it make it easier for the family to visit less?
And who owns the information? This worries me a lot about a lot of the latest tech coming out.
A companion that knows when someone wakes up, what they talk about, whether they took their medication, how often they move around the house and perhaps what they are worried about is sitting on top of an extraordinarily intimate dataset.
The convenience is obvious but so is the privacy risk.
There are more immediate questions too. Who is responsible if an AI reminder fails or gives the wrong information? What happens when the person using it has dementia or another form of cognitive impairment? How does a family monitor an elderly parent without turning a tool designed to support independence into a form of surveillance? I didn't find much on this.
There is also the question of money.
The simplest voice companions are relatively accessible compared with a humanoid robot costing hundreds of thousands of rands (if they ever make it to our shores). But even the cheaper systems depend on reliable internet connectivity and ongoing services.
Language is another consideration in a country with 11 official languages. A system that works beautifully in English may be far less useful to an older person who is more comfortable communicating in another language.
What this could mean for South African families
South Africa has not yet emerged as a major testing ground for the sophisticated elder-care robotics programmes being rolled out in Japan and China.
But it is facing the same pressures driving those programmes: a growing elderly population, families separated by distance and a shortage of people available to provide care.
That does not mean a robot can replace a daughter, a son, a nurse or a caregiver. But perhaps it can help those people do a better job, and help an older person remain independent for longer.
The technology is likely to move from the voice on the table to machines that can move through the house, carry things, fetch objects, provide physical assistance and eventually help with some of the most exhausting tasks of caring for another human being.
The first generation of elder-care robots may simply keep someone company, the next may help them get through the day and the generation after that could be helping them get out of bed.
For a world running short of people to care for its elderly, that future is no longer some futuristic movie. The robots for this world are already being built, and for my sister and I, we hope it will happens in our lifetime. And for those with ageing parents, in their lifetime.
vivian.warby@nationalmg.co.za