The four-day week the Chancellor didn't mention

Start with a number that should be more famous than it is. A British worker, on average, produces meaningfully less in an hour than a French, German or American one — the productivity gap that has shadowed this country for the better part of two decades and that every chancellor, including the current one, swears to close. A year ago, in the 2025 Spending Review, Rachel Reeves placed her bet on how: roughly £2 billion on artificial intelligence, on the theory that machines will finally let British workers produce more per hour. This week's hiring data is the first real evidence of that bet's human edge arriving. Set aside whether the gamble pays. Ask instead the question that the whole conversation steps around: if it works — if the same output starts taking less human time — what happens to the time?

There are only two honest answers, and a society chooses between them. Either the gains are taken as more output — the same hours, more produced, more sold, more profit — or they are taken as less time — the same output, fewer hours, the surplus handed back to the human being as life. For two hundred years of industrialisation we mostly chose the first, with occasional, hard-won exceptions: the weekend, the eight-hour day, the bank holiday, each one a moment when a society decided that a slice of the productivity dividend should be returned as time rather than things. We are, whether the Chancellor said so or not, standing at another such moment.

For most of its life, the four-day week was an idea you weren't allowed to take seriously — a utopian's indulgence, fine for a Scandinavian pilot but unthinkable for a real economy. That defence has quietly collapsed. The large UK trials of recent years, in which firms moved to four days with no loss of pay, produced results that embarrassed the sceptics: the overwhelming majority of participating companies chose to keep the arrangement, revenue broadly held, and the human indicators — burnout, sick days, resignations — moved sharply in the right direction. The finding that mattered most was the dullest: for a great many knowledge jobs, the fifth day had not been producing very much. It was attendance, not output.

Now layer AI on top of that. If a marketing team or a legal department or an accounts function genuinely becomes a third more productive with the tools the Chancellor is funding, the fifth day becomes not just unproductive but visibly, measurably surplus. The machine doesn't create the case for shorter hours; it removes the last excuse against it. Which is precisely why the silence is so striking. The same government spending billions to compress the work has no position whatsoever on what to do with the time that compression frees.

The reasons for the silence are not mysterious, and they are worth naming plainly. A government that championed a four-day week would be accused, within a news cycle, of wanting Britain to work less while its rivals work more — a caricature, but a potent one in a country with a deep moral attachment to effort. There is the genuine complication that shorter hours are far easier to imagine for a salaried office worker than for a nurse, a carer, a delivery driver, or anyone whose output is their physical presence — and a reform that freed the comfortable while leaving the frontline untouched would deserve the backlash it got. And there is the simple fact that the dividend, if it comes, lands first as private profit on a company's balance sheet, not as a public good a minister can hand out. To turn productivity into time requires either that firms choose to, or that the state makes them — and neither is a fight this government has shown any appetite to pick.

So the question gets left to individual employers, which means it gets answered by power. The firms already running four-day weeks are mostly those that could afford the experiment and wanted the recruiting edge. The worker most likely to see AI's productivity gain returned as free time is the one who already had the most leverage; the worker most likely to see it returned as a P45 is the one who had the least. Left ungoverned, the most humane possibility of the AI transition — that it gives us our time back — will distribute itself exactly as unequally as everything else does. That is not a prediction. It is what's already happening, quietly, firm by firm.

This is where the founding question of this publication stops being abstract. "What do humans do when we don't have to work?" sounds like a thought experiment for a distant automated future. It is not. It is a live policy choice the country is making right now, by default, by declining to make it. Every hour of productivity that AI unlocks over the next decade will go somewhere — into output or into time, into profit or into life, to the powerful or to everyone. The Spending Review funded the machine that creates the surplus and said nothing about who receives it.

The four-day week is not the whole answer, and it may not even be the right vehicle; perhaps the dividend is better taken as higher pay, or shorter careers, or longer leave, or a serious shortening of the working year rather than the week. The vehicle is arguable. What is not arguable is that a choice is being made, and that the absence of a national conversation about it is itself a decision — the decision to let the gains fall where power already sits. Britain has handed back the productivity dividend as time before. The weekend was not a gift of nature; it was won. The interesting thing about this moment is that the surplus is about to be larger than at any point since, and the country, so far, has decided not to discuss what to do with it.

So the Long Table's question for the week is the one the Chancellor's billions raise and her speech avoided. We are about to get more time. The only thing left to decide is whose it is. That is worth more than a footnote in a Spending Review — and we suspect it will, eventually, be the whole argument.

◆ Why it matters here — The same £2bn AI bet that is thinning entry-level hiring is also, if it works, large enough to be taken partly as time rather than output. Britain has no national position on which — so by default the dividend goes to whoever already holds the most leverage.