Can robots care for the elderly? Scientists race to solve aging care crisis

In a laboratory northwest of London, three black metal hands move with deliberate slowness across an engineering bench. Each one has four fingers and a thumb, articulated in a way that mimics human anatomy almost perfectly. Rich Walker, who directs the company that built them, laughs when asked if this is the beginning of some dystopian future. "We're not trying to create a Terminator," he says. What Walker wants instead is something more modest and more urgent: a robot that can sweep a floor, prepare a meal, help an elderly person move from bed to wheelchair. The ambition, though, extends far beyond household chores. Across the developed world, researchers and companies are betting that machines like these could help solve one of the defining challenges of the coming decades—how to care for rapidly aging populations when there simply aren't enough human caregivers to go around.

The pressure is most acute in places like Japan and the United Kingdom, where populations are graying faster than anywhere else. A decade ago, the Japanese government began subsidizing robot manufacturers to develop and deploy these machines in nursing homes. The logic was straightforward: fewer young people to do the work, more elderly people needing care. Why not let machines fill the gap? James Wright, an artificial intelligence researcher at Queen Mary University in London, spent seven months observing this experiment firsthand. He watched three different robots in action. The HUG, developed by Fuji Corporation, looked like an extremely sophisticated walker with padded supports that let caregivers safely transfer residents from bed to chair. The Paro resembled a small seal and was designed to stimulate dementia patients through touch and sound. The Pepper was a small humanoid robot that could lead exercise classes by demonstrating movements with its arms.

What Wright expected to find was eager adoption by overwhelmed caregivers. What he actually discovered was almost the opposite. The robots, it turned out, created more work than they saved. Caregivers spent enormous amounts of time cleaning and recharging them, troubleshooting when something went wrong. After a few weeks, staff concluded the machines were a net burden. The HUG had to be constantly moved out of residents' way. The Paro caused emotional distress when one resident became too attached to it. The Pepper was too short for everyone to see and its high-pitched voice was hard to understand. Within weeks, the robots were being used less and less. The manufacturers responded with their own defenses—the HUG's designers said they'd made it more compact, Paro's creator pointed to studies showing therapeutic benefits, Pepper's new owners claimed significant software improvements. But the fundamental problem remained: the gap between what robots could do in theory and what they could actually accomplish in a real care facility.

Yet the push continues. Praminda Caleb-Solly, a professor at the University of Nottingham, is determined to make this work. She created a network called Emergence that connects robot manufacturers directly with elderly people and care providers to understand what they actually need. The answers vary. Some want voice interaction and non-threatening appearance. Others prefer cute designs. But the most consistent request is brutally practical: the robot should charge itself, clean itself, and adapt to changing needs. "We don't want to care for the robot," one person said. "We want the robot to care for us." In Cheltenham, a home care provider called Caremark has been testing a voice-activated robot called Genie with some of its clients. One man with early-onset dementia enjoys asking it to play Glenn Miller songs. The reactions have been mixed—some people loved it, others were indifferent. The company's director, Michael Folkes, is careful to say the goal isn't replacement. "We're trying to build a future where caregivers have more time to care."

The technical challenges are immense. For a robot to be truly useful in a care setting, it needs human-like dexterity. Walker's robotic hand, equipped with 100 sensors, can solve a Rubik's cube with one hand and make an "OK" gesture with precision. But it still can't use scissors or pick up fragile objects. The subtlety of human touch—the way we adjust pressure and feel feedback—remains elusive. Walker and 35 other engineering firms are part of what's called the Robotic Dexterity Programme, funded by the UK's Advanced Research and Invention Agency. They're studying how animals move to redesign robots from the ground up. Jenny Read, who leads the project, notes that evolution has produced grace and efficiency in animal bodies. "Grace," she says, "is really a form of efficiency." Meanwhile, a Danish engineer named Guggi Kofod is developing artificial muscles made from soft material that stretches and contracts when electric current is applied, mimicking real muscle. His company, Pliantics, is working with Shadow Robot to create a human-sized robotic hand with artificial muscles that could detect minute pressure changes and know exactly when to stop squeezing. Kofod's motivation is personal—people close to him have died from dementia, and he wants to build systems that help caregivers and allow patients to maintain dignity.

But beneath the technical optimism lies a harder question about what this future actually looks like. James Wright worries that if robots become widespread, the economics could push care facilities toward larger, standardized operations where robots do most of the work and human caregivers are paid minimum wage just to support the machines. The vision of robots giving caregivers more time for meaningful interaction could invert into something darker. Gopal Ramchurn, an AI professor at the University of Southampton and CEO of Responsible AI, takes a different view. He sees an enormous industry emerging to meet the genuine shortage of caregivers as populations age. But he also points to Elon Musk's Optimus robot, which served drinks at a Tesla event, as a sign that whether we're ready or not, robots are coming. "We're trying to anticipate this future before large tech companies deploy these machines without asking what we think," he says. The real work now, he argues, is developing regulations that ensure robots serve human needs rather than replacing human connection. The question isn't whether robots will arrive in care facilities. It's whether we'll be prepared to shape what that arrival actually means.