In a Friday commentary, Professor Dame Henrietta Moore argued that it is time to rethink welfare as AI reshapes the economy and wears down the brains of some of the most vulnerable workers. Moore, founder and director of the UCL Institute for Global Prosperity and co-leads The New Alignment, the institute’s G20 policy programme, said governments should ask how every fiscal policy will affect the cognitive load of the poorest people it touches and should provide universal basic services free at the point of use. AI, she said, is adding to the cognitive load of workers already stretched thin by unpredictable pay, insecure housing and mounting debt, and the lowest paid feel it most. Retraining will be harder, she argued, for workers whose minds are preoccupied with worries over pay, housing and debt.

Moore cited a 2024 analysis of 112,000 people in the Journal of Economic Psychology to argue that financial pressure constrains thinking. The analysis found that financial scarcity measurably impairs how well people think and that the effect lifts when financial pressure does too. Among low-paid workers in rural India who received their wages a few days early, productivity was 7% higher and mistakes were fewer. In South Africa, after the pension age for men was lowered from 65 to 60, those who became eligible earlier showed slower memory decline and were about five percentage points less likely to develop dementia.

Those pressures can accumulate, Moore says. Livelihood pressures stack up to exceed what one mind can handle, and poverty manufactures its own interest payments. Her hypothetical week begins with a worker’s rent going up and her losing a shift at the last minute, something common for workers without guaranteed hours. She puts the shortfall on a credit card. Preoccupied with finding the rent, she returns a benefits form late and loses part of her benefits payment as a penalty. Moore says the three setbacks leave the household poorer in money, headspace and health.

The workplace changes Moore describes can arrive before jobs disappear. Automation makes work less predictable, shifting and shortening schedules long before it eliminates jobs, and she cites Walmart staff dealing with AI-set schedules as an example. Research from the Boston Consulting Group found that overseeing AI tools adds mental fatigue. Her warning is direct: “Workers required to retrain for the AI economy will struggle to learn new skills if their minds are already occupied by worries over pay, housing and debt.”

Moore places the change in the AI economy within a broader argument about economics. She says the human brain, which directs labour, judgment and resilience, has rarely been considered an asset in its own right, but the emerging AI economy is changing that. As machines take on routine work, what becomes most valuable is the judgment, creativity and insights only people can provide. Economists coined a term for this in recent years: brain capital. The World Economic Forum, cited by Moore, finds that analytical thinking is now the skill employers value most and estimates that 59 in every 100 workers will need training to adapt to the AI shift by 2030.

Moore also argues that current measures miss a central condition of brain health. Two new indices rank countries and US states on brain capital, she says, using about 60 indicators that range from child poverty to the prevalence of Parkinson’s disease. Neither measures whether people’s livelihoods are secure, including whether income is predictable, debts manageable and homes safe. The World Health Organization counts safety and security among foundations of brain health, Moore notes, but the index builders did not ask what drains people’s mental capacity.

At the first Global Brain Economy Summit in New York in September, Moore framed the problem this way: “How much room is there for creativity and learning when one’s mind is consumed by anxieties about paying rent or for food to last through the week?” She calls the missing liability “cognitive taxation” — a tax no government voted for, levied on attention, memory and the capacity to learn, and collected through rents, debts and welfare paperwork. Worry over insecurity runs in the background, she writes, like an app draining a phone battery; over years, chronic stress wears down the brain itself, and AI is raising the rate.

Britain shows what insecurity already costs before AI adds to it, Moore writes. About 2.7 million people there are out of work through long-term sickness, she says. Research from Moore’s institute found that people facing insecurity across several areas of life carry an earnings penalty as large as the gender pay gap, a penalty that roughly doubles after seven years. Moore says those findings suggest that worry erodes how well people work. The UK’s AI minister said the government must seriously consider that the technology could have “an unprecedented impact” on the jobs market. She sums up: “Worries abound and compound.”

Moore also questions cash alone as an answer to job losses. When Elon Musk told reporters at the White House in September that AI would bring “universal high income,” he offered, in Moore’s framing, Silicon Valley’s favourite answer: cash. A study funded by Sam Altman that gave low-income Americans $1,000 a month found that mental health gains faded after the first year. Extra cash helps, Moore says, but it cannot guarantee a child a nursery place or a bus that gets someone to work on time. These pressures already tax the minds of people in work, she says, and no payment removes them.

Her alternative is universal basic services: childcare, housing, transport and digital access, free at the point of use. Moore compares the principle with the one behind the UK’s National Health Service, under which people can rely on essential support without facing a bill when they use it. Extending free childcare alone, she says, is expected to bring 60,000 non-working parents into jobs in the UK.

Moore wants every government to ask how each fiscal policy will affect the cognitive load of the poorest people it touches. Britain, which takes over the G20 presidency after December’s summit in Miami, should put that question to the world’s largest economies, she says. She wants the test included in treasury fiscal plans, brain-capital indices and AI-powered public services before they launch.

Moore closes by comparing the present challenge with environmental accounting. Sixty years ago, she writes, societies learned to count the damage economies do to the air and water, helping change how infrastructure was built for the better. Her conclusion: “Now that AI is making our minds the most valuable thing we have, it is time we count the damage done to them.”