Thursday, 12 June 2025

A quick look at how AI is helping in social care

Britain | AI and social care

Britain’s AI-care revolution isn’t flashy—but it is the future

To see the future of social care, come to England’s Black Country

Illustration: Eiko Ojala

Jun 5th 2025|Dudley, West Midlands

Listen to this story

ARTIFICIAL INTELLIGENCE (AI) is typically associated with Silicon Valley coders or researchers in Shanghai, not chain-smoking care workers in the Black Country. Yet in England’s post-industrial heartland Samantha Woodward, a manager at Cera, a home-care company, arranges carers’ schedules with Amazon-like efficiency. A custom-built app plots the quickest routes to see the most clients in the shortest time. Carers log their arrival by sharing their GPS location. Alerts ensure that medication is given on schedule. So good are the data that Cera can even predict which workers are likely to quit (staff turnover is said to be down by 20%).

England may seem an unlikely pioneer in an AI-care revolution. Whereas Japan is deploying robots, a quarter of care providers in England still keep paper records. The country’s care system is underfunded and overstretched. Yet necessity is spurring innovation, albeit not from cash-strapped local councils (over 40% of their spending on services already goes on adult social care). Young entrepreneurs are teaching carers to use tools more common to delivery services and dating apps.

One of them is Ben Maruthappu, Cera’s founder. He launched Cera after failing to find reliable care for his mother (an origin story typical of aspiring age-tech entrepreneurs). If groceries could be tracked in real time to the doorstep, he reasoned, then why not medication? Today Cera claims to have created Europe’s largest home-care data set—over 200bn data points—to train AI that predicts patients’ needs.

A promising use of AI is to predict falls. Falls are among the gravest risks for the elderly: hip fractures are their most common cause of accidental death. They also cost the National Health Service (NHS), which is separate to social care, around £2bn ($2.7bn) a year. With its app, which uses algorithms to predict fall risk, Cera claims to have cut falls by a fifth. A peer-reviewed study from 2022 found that its app had reduced hospitalisations by 52%. In March the NHS said it would work with Cera to roll out its AI tool across the country.

Fifty kilometres south of Ms Westwood’s patch is The Lawns, a nursing home. There, another tech adopter, Melanie Dawson, a former rugby player turned care manager, has overseen an NHS pilot using acoustic-monitoring devices. White boxes combine motion sensors and machine learning to detect unusual movements or noises in the residents’ rooms, a kind of Shazam for ambient noise. Over a year-long trial, falls decreased by 66%, and staff made 61% fewer checks in person. With fewer disturbances, residents also slept better and, with less daytime-napping, ate more. When the pilot finished, the home chose to keep using them at its own cost.

At Cavendish Park, in Worcestershire, residents can play with robotic companion pets and try an interactive driving simulator. Most, though, stay in their rooms, chatting to smart speakers called Alexa (made by Amazon). Cavendish Park was the first home in Europe to build Alexas, adapted to care homes, into its infrastructure, integrating them with its alarm and communication systems. Residents can use them to ask for help, for a drink or to make an emergency call.

AI’s adopters say the tech is meant to support staff, not replace them. At Cavendish Park, carers scan residents’ faces using PainChek, a mobile app, which detects micro-expressions linked to pain. Within seconds, they can detect pain more quickly and accurately than a trained nurse.

Yet in other cases, AI appears to be filling gaps left by a shortage of carers. In Northamptonshire, Cera is testing robots to give routine prompts, like whether clients have eaten or taken their medication. Though fine for the most mundane tasks, they are no replacement for Ms Westwood’s kind and competent team. At Cavendish Park, residents treat their chatbots as companions (one 99-year-old says she spends all her time playing “Quick-fire Quiz”). Meryem Tom, who leads the Alexa division at Amazon, insists that the Alexas are “complementary” to humans.

One risk is that care workers could become tethered to digital metrics, a grim prospect for such a human job. Consent is another concern. Many of the clients whom AI might best help have dementia, raising doubts about whether they can meaningfully opt in. Some homes already feel like prisons: the dementia wing at Cavendish Park, Chatsworth House, is locked with keypads. AI could worsen that controlling urge. With enough surveillance, warns Andrew Sixsmith, a gerontology professor at Simon Fraser University, they risk resembling Jeremy Bentham’s Panopticon, a theoretical prison where a single guard can watch every cell.

Still, for those needing care, the benefits of safety and independence appear to outweigh the risks—at least for now. Brenda Adkin is a 101-year-old Cera client who recently suffered a fall. Still enjoying a sherry with her neighbour in the morning, she has no wish to go into a care home. “I like my independence,” she says. AI helps the carers she loves keep her at home. Solving the care crisis will take more of this kind of innovation, not less. ■

For more expert analysis of the biggest stories in Britain, sign up to Blighty, our weekly subscriber-only newsletter.

Trump, tariffs and manufacturing jobs in the US

Finance & economics | Blue-collar blues

Factory work is overrated. Here are the jobs of the future

America is trapped by its industrial fantasies

Illustration of a conveyor belt with parcels on with a cut in the middle in the shape of a worker with a hard hat on — Illustration: Alex Nabaum

Jun 10th 2025

Trumpian types are unanimous: America needs factories. The president describes how workers have “watched in anguish as foreign leaders have stolen our jobs, foreign cheaters have ransacked our factories and foreign scavengers have torn apart our once beautiful American dream”. Peter Navarro, his trade adviser, says that tariffs will “fill up all of the half-empty factories”. Howard Lutnick, the commerce secretary, offers the most cartoonish pitch of all: “The army of millions and millions of human beings screwing in little screws to make iPhones—that kind of thing is going to come to America.”

For years, politicians and some economists have linked manufacturing’s long decline to stagnant wages, hollowed-out towns and even the opioid crisis. In the 2000s alone America shed nearly 6m factory jobs. Such work often offered high-school leavers a route to a stable, quietly prosperous life. It sustained entire cities, earning Pittsburgh the moniker “Steel City” and Akron that of “Rubber Capital of the World”. Little surprise, then, that politicians across the spectrum want the jobs back. Indeed, President Joe Biden shared the same dream as his successor, even if he hoped to achieve it by different means. “Where the hell is it written”, he asked, “that we’re not going to be the manufacturing capital of the world again?”

Yet there is a problem: even if industry returns, the old jobs will not. Manufacturing produces more than in the past with fewer hands—a transformation much like that undergone by agriculture. Accessible, middle-class work of the sort that once drew crowds to the factory gates in America’s Fordist heyday has all but vanished. According to our analysis, the most similar work to the manufacturing jobs of the 1970s is not to be found in factories, which are now automated and capital-intensive, but in employment as an electrician, mechanic or police officer. All offer decent wages to those lacking a degree.

Whereas almost a quarter of American workers were employed in manufacturing in the 1970s, today less than one in ten is. Moreover, half of “manufacturing” jobs are in support roles such as human relations and marketing, or professional ones such as design and engineering. Fewer than 4% of American workers actually toil on a factory floor. America is not unique. Even Germany, Japan and South Korea, which run large trade surpluses in manufactured goods, have seen steady falls in the share of such employment. China shed nearly 20m factory jobs from 2013 to 2020—more than the entire American manufacturing workforce. Research from the IMF calls this trend “the natural outcome of successful economic development”.

As countries grow richer, automation raises output per worker, consumption shifts from goods to services, and labour-intensive production moves abroad. But this does not mean factory output collapses. In real terms, America’s is over twice as high as in the early 1980s; the country churns out more goods than Japan, Germany and South Korea combined. As the Cato Institute, a think-tank, points out, America’s factories would, on their own, rank as the world’s eighth-largest economy.

Even a heroic reshoring effort eliminating America’s $1.2trn goods-trade deficit would do little for jobs. In the production of that amount of goods, about $630bn of value-added would come from manufacturing (with the rest attributable to raw materials, transport and so on). Robert Lawrence of Harvard University estimates that, with each manufacturing worker generating around $230,000 in value added, bringing back enough production to close the deficit would create around 3m jobs, half on the factory floor. That would lift the share of the workforce in manufacturing production by barely a percentage point. Assume this was achieved by levying an average effective tariff rate of 20% on America’s $3trn of imports, and it could push up prices by around $600bn, or $200,000 per manufacturing job “saved”.

It is a high price for jobs that are not as attractive as in the past. Seven decades ago, factories offered a rare bundle: good pay, job security, union protection, plentiful employment and no degree requirement. By the 1980s manufacturing workers still earned 10% more than comparable peers in other parts of the economy. Their productivity was also growing faster.

Today factory-floor work lags behind non-supervisory roles in services on hourly pay. There has also been a collapse in the manufacturing wage premium, which compares earnings for similar workers by controlling for age, gender, race and more. Using methods similar to the Department of Commerce and the Economic Policy Institute, we estimate by 2024 the premium had more than halved since the 1980s. For those without a college education, it has gone entirely, even though such workers still enjoy a premium in the construction and transport industries. Productivity growth has fallen, too: output per industrial worker is now growing more slowly than per service-sector worker, suggesting wage growth will be weak as well. A crucial component of the “manufacturing jobs are good jobs” argument no longer holds.

And a job in industry is harder to attain, too. Modern factories are high-tech, run by engineers and technicians. In the early 1980s blue-collar assemblers, machine operators and repair workers made up more than half of the manufacturing workforce. Today they account for less than a third. White-collar professionals outnumber blue-collar factory-floor workers by a wide margin. Even once obtained, a factory job is far less likely to be unionised than in previous decades, with membership having fallen from one in four workers in the 1980s to less than one in ten today.

To find the modern equivalent of such jobs, we looked for employment with the same traits. What offers decent pay, unionisation, requires no degree and can soak up the male workforce? The result: mechanics, repair technicians, security workers and the skilled trades.

Over 7m Americans work as carpenters, electricians, solar-panel installers and in other such trades; almost all are male and lack a degree. The median wage is a solid $25 an hour, unionisation is above average and demand is expected to rise as America upgrades its infrastructure. Another 5m toil as repair and maintenance workers—think HVAC technicians and telecom installers—and mechanics, earning wages well above the factory-floor average. Emergency and security workers also show similarities; over a third are union members.

Still, these jobs differ from manufacturing in one way: there is no such thing as an HVAC company town. Factories once powered cities, creating demand for suppliers, logistics and dive bars. The new jobs are more dispersed and, as such, less likely to prop up local economies. Yet, although the benefits are diffuse, they are almost as large. Nearly as many people are employed in such categories as held manufacturing jobs in the 1990s. With better wages, less credentialism and stronger unions, they look more attractive than modern factory jobs to working-class Americans.

The future is drifting even further from factories. Skilled trades and repair workers should see growth of 5% over the next decade, according to official projections; the number of manufacturing jobs is expected to fall. The fastest-growing categories for workers without degrees are in health-care support and personal care, which are expected to grow by 15% and 6%, respectively. These include roles such as nursing assistants and child-care workers, and do not look anything like old manufacturing jobs owing to their low pay. The task, as Dani Rodrik of Harvard puts it, is to boost the productivity of the jobs that are actually growing. Perhaps that might include ensuring the adoption of AI, whether for managing medication or diagnosis.

In the late 18th century, Thomas Jefferson viewed farming as the foundation of a self-reliant republic. Influenced by French physiocrats who saw agriculture as the noblest source of national wealth, he believed that working the land was the path to liberty and abundance. By the 20th century, factory work had inherited that symbolic role. But like farming before it, manufacturing employment fades with rising prosperity and productivity. The heart of working-class America now beats elsewhere. ■

Quote of the day