For you
The full edition
£550+ VAT
The complete report, to download and keep.
- The full report to download
- Full access online
The Future Collective · Flagship report · October 2026
When intelligence gets cheap, what gets expensive?
What happens to work, money, health, energy, food, learning, truth and power over the decade to 2035. Twelve questions, four possible worlds, and twelve dated calls we will mark in public every October.
What it covers
One chapter for each part of life, from work and money to food, truth and power. Each asks one question about 2035, answers it, and makes a call we will mark in public every October.
Ways in
For you
£550+ VAT
The complete report, to download and keep.
For your leadership team
£3,500+ VAT · half day
We present the report to your team, then help you work out the so what: what it actually means for your company.
Or email matt@future-collective.co.uk
For your leadership team
£8,000+ VAT · full day
We take the parts of the report that matter most to your company and industry, and run a full-day workshop with your team.
Or email matt@future-collective.co.uk
In five minutes
2035 is as far ahead of us as 2017 is behind. Much of it is already settled: the people who will be working, retiring and starting school then are alive now, and many of the laws and building projects that will shape it already have dates. This report starts from what is settled, says what is likely and how sure we are, and then makes twelve calls we will mark in public every October.
Machine intelligence is getting cheaper at a remarkable rate. When one thing becomes abundant, value moves to whatever is still scarce beside it. Across the twelve chapters the same four things keep running short.
The caveat runs through every chapter: cheap is not the same as useful. The gap between what the machines can do and what organisations actually do with them is wide.
Read together, most answers say that something people fear or hope for does not arrive on time. A few say a slow trend keeps going. Only three say something moves fast: the machines' appetite for electricity, the share of it from renewables, and the number of people who get their news from a machine. That is the argument in twelve lines. Intelligence is getting cheap fast. Almost everything it needs in order to matter is scarce and slow.
Two things decide which 2035 we get: whether people and institutions learn to use cheap intelligence widely, and whether the physical and public things get built and paid for. That gives four worlds, which we follow through one invented person, Sam, aged 27 in Stockport. In The Long Grind, nothing collapses and everything takes slightly longer than it should. In The Crossing, institutions redesign how they think and the gains reach people like Sam. In The Split, the remarkable things all happen, to someone else. In The Shock, something large from outside stops the money and materials everything else assumed. Our view is that Britain is in The Long Grind today, drifting towards The Split. The Crossing is available, and it is mostly a matter of design and delivery.
Every call names the official figure that will settle it. Our confidence levels imply we should get about eight of the twelve right: all twelve would mean we were too timid, five that we were overconfident. Before publishing we put the calls to a panel of 24 synthetic forecasters. It thought two of our calls too easy, so we made them harder: unemployment below 5.5% rather than 6%, and fewer than 75% of NHS patients seen within 18 weeks rather than 92%. Where it thought us too sure, on farming and university, we have held and said so. The scorecard comes every October.
Our confidence, and where a synthetic forecasting panel put the same call

Part one · The argument
Proof, people, physical capacity and coherence.
Photograph: Unsplash
Read the chapterPart one
2035 is exactly as far ahead of us as 2017 is behind. Nine years. Most of us can remember 2017 without effort, which is a useful corrective to the idea that the mid-2030s is science fiction.
A surprising amount of 2035 already exists. The eighteen-year-olds of that year are in primary school now. Everyone who will draw a state pension in 2035 has been alive for more than half a century. The children who will start school in England and Wales in 2029 and 2030 were born last year, and there were fewer of them than in any year since 1977 (ONS). The grid connections, reservoirs and railway lines that will be finished by then have, for the most part, already been applied for. Several have already been delayed.
So this is not a book of guesses. It starts with what is already settled: the laws with dates on them, the people already born, the contracts already signed. Then it moves to what is likely, and says how sure we are. Then it stops at the questions nobody can answer yet, and shows the two ways each could go.
We have done this for twelve parts of life, from work and money to food and power. And we finish by doing something forecasters generally avoid. We make twelve specific calls about 2035, write down what would prove each one wrong, and promise to mark our own homework in public every year.
A report about everything is an absurd thing to attempt. We know. But most people do not live in a sector. They live in a house, with a job, a body, a bank account and a view about the news. The future arrives in all of those at once, and it seemed worth trying to describe it the same way.
Full edition
The rest of this part is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

Part two · Twelve domains
Twelve chapters, one for each part of a life.
Photograph: Unsplash
Read the chapterPart two
Twelve chapters follow, one for each part of a life. Each stands alone and takes about ten minutes.
| Chapter | The open question | |
|---|---|---|
| 01 | Work | Is the bottom rung of the career ladder gone for good? |
| 02 | Machines | Does capability turn into transformation, or does adoption lag by a decade? |
| 03 | Money | Do ageing states fix their finances by choice, or have it forced on them? |
| 04 | Health | Does new medicine bend the illness curve, or widen the gap? |
| 05 | Energy and climate | Does clean power cut fossil fuel use, or only meet new demand? |
| 06 | Buying | Does the AI agent become the buyer, or stay the adviser? |
| 07 | Making and moving | Does the West hold its electric car deadlines, and whose factories supply the cars? |
| 08 | Places | Does Britain build the homes, or legislate the ambition? |
| 09 | Food | Does food stay cheap? |
| 10 | Learning | Does AI raise what people can do unaided, or only what they can produce with it? |
| 11 | Truth | Can verified reporting survive when a machine stands between reader and source? |
| 12 | Power | Managed rivalry between armed blocs, or something worse? |

01 · Work
No. Not enough to break the jobs market.
Photograph: Unsplash
Read the chapter01 · Work
The robots are not coming for your job. The awkward truth is closer to the opposite: over the next decade there will not be enough workers, and the people struggling most will be the ones trying to get their first job.
01/12
01 · Work
20 of 26
years in which UK unemployment has stayed below 6%.
Source: ONS · Landscape drawn from the chapter’s data
Fewer workers per pensioner. From mid-2026 the UK is projected to record more deaths than births, so every bit of population growth comes from migration. By 2034 the number of people of pension age rises by 1.8 million to 14.2 million, while the working-age population grows by only 1.5 million (ONS). Across the rich world the working-age population shrinks about 8% by 2060, and in South Korea by up to 46% (OECD).
A longer working life. The UK state pension age reaches 67 by 2028. A review due by March 2029 will decide whether 68 comes sooner than the 2044 to 2046 now in law. China began raising its retirement ages in January 2025 (NPC Observer).
New rights at work. The Employment Rights Act 2025 is arriving in stages. Day-one sick pay began in April 2026. From January 2027, protection against unfair dismissal starts after six months, not two years, with guaranteed-hours offers for zero-hours workers to follow (Lewis Silkin tracker).
Slow growth and a reshuffle, not a collapse. The US Bureau of Labor Statistics expects employment to grow 3.5% between 2025 and 2035, against 10.9% in the decade before. Healthcare supplies more than a third of the new jobs. Office and administrative support shrinks by about 750,000 (BLS).
Skills that expire faster. Employers surveyed by the World Economic Forum expect 39% of workers' current skills to change or go out of date by 2030. Of every 100 workers, 59 will need training, and 11 are unlikely to get it.
Exposure almost everywhere. One in four workers worldwide is in an occupation with some exposure to generative AI, though the ILO expects most of those jobs to change, not vanish.
A modest official forecast. The Office for Budget Responsibility expects UK productivity to grow 1% a year and unemployment to fall back to about 4% by 2030. It has also modelled a version in which technology pushes structural unemployment up to 5.5%. Even its downside scenario for the economy peaks at 6.8% in 2026-27.
Is the bottom rung of the career ladder gone for good?
The case that it is. Payroll data analysed at Stanford shows employment of 22 to 25 year olds in the jobs most exposed to AI running about 19% below where it would otherwise be, as of June 2026. A year earlier the gap was 15%. The pattern looks like less hiring, not more sacking, and the effect is concentrated where AI does the task outright. Experienced workers in the same jobs show no comparable gap. In the UK there are now 140 applications for every graduate vacancy, up from 86 two years earlier (Institute of Student Employers).
The case that it is a dip. The Stanford authors say their pattern is descriptive, and some of it predates ChatGPT. The Yale Budget Lab finds no clear link between AI exposure and unemployment. Danish administrative data rule out effects on pay or hours bigger than 2%, two years in (Humlum and Vestergaard).
Why it matters: junior jobs are where judgement gets built. An organisation that stops hiring 23-year-olds in 2026 discovers in 2034 that it has no 31-year-olds who know how anything works. That is a slow spending-down of cognitive capital, and it will not show up in this year's accounts.
The productivity promise. In a randomised trial, experienced developers using AI took 19% longer while believing they were 20% faster (METR). A follow-up could not establish a speed-up either. Nobody should trust a self-reported AI saving, including their own.
Work gets heavier, not lighter. Researchers at Berkeley found AI did not reduce workloads; people worked faster, wider and longer (Harvard Business Review). In a survey by Upwork, a freelance platform, 88% of the most productive AI users reported burnout (Upwork). And 41% of workers say they have received what researchers call "workslop": polished-looking AI output that takes nearly two hours to fix.
Sickness. 2.78 million people in the UK are out of the workforce through long-term illness (ONS). The government warns that a further 600,000 could leave the workforce by 2030 if current trends continue (UK Parliament).
The young. Nearly a million 16 to 24 year olds, 13%, are not in education, employment or training (ONS).
Migration as a fix. UK net migration was 171,000 in 2025, down 82% from its peak (ONS). Japan, facing a projected shortfall of 3.84 million workers by 2035 (Persol and Chuo University, via Nippon.com), now has a record 2.57 million foreign workers (Japan labour ministry, via Nippon.com). Rich countries are about to compete for the same people.
A class divide in flexibility. Hybrid working reaches 45% of people earning over £50,000 and 8% of those under £20,000 (ONS).
If you lead an organisation
If you are simply trying to stay employed

02 · Machines
Yes.
Photograph: Unsplash
Read the chapter02 · Machines
The machines are improving faster than almost anyone predicted and being put to use more slowly than almost anyone admits. Both things are true, and the distance between them is the story of the decade.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

03 · Money
No. A bigger benefits bill instead.
Photograph: Unsplash
Read the chapter03 · Money
Money is getting faster, more digital and easier to steal. Underneath that, something slower and larger is happening: ageing countries have made promises they have not worked out how to pay for.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

04 · Health
It survives, but misses its own promise.
Photograph: Unsplash
Read the chapter04 · Health
Medicine is about to have one of the best decades in its history. Whether that makes the country healthier is a different question, and the answer depends less on the laboratory than on the postcode.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

05 · Energy and climate
Only just.
Photograph: Unsplash
Read the chapter05 · Energy and climate
Two things are now true at once. Clean energy is winning faster than its critics said it could. And the world will pass 1.5 degrees of warming anyway. The 2030s are about living with both.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

06 · Buying
We already have, per head.
Photograph: Unsplash
Read the chapter06 · Buying
Shopping has been declared dead, reborn and revolutionised so many times that it is worth starting with a dull fact. In Britain, roughly £73 of every £100 spent in retail is still spent in a shop (ONS).
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

07 · Making and moving
No.
Photograph: Unsplash
Read the chapter07 · Making and moving
The West spent a decade promising to bring its factories home. China spent the same decade building them. This chapter is about what follows from that, and about how we will get around while it plays out.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

08 · Places
No.
Photograph: Unsplash
Read the chapter08 · Places
Buildings are the slowest technology we have. Most of the homes, streets and offices of 2035 are standing today, which makes this the easiest chapter to predict and the hardest in which to change anything.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

09 · Food
Yes. Fewer farms, same share of food.
Photograph: Unsplash
Read the chapter09 · Food
The future of food was supposed to arrive in a laboratory dish. So far it has arrived as a weather forecast, a weight-loss injection and a slightly more expensive bar of chocolate.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

10 · Learning
For most, yes. Not for all.
Photograph: Unsplash
Read the chapter10 · Learning
Education is being hit by two things at once: a machine that can do the homework, and a shortage of children to set it for. The second gets far less attention and is far more certain.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

11 · Truth
Yes, nearly a third of us.
Photograph: Unsplash
Read the chapter11 · Truth
The fear was that we would drown in convincing lies. What is happening is quieter and in some ways stranger: the people who check facts for a living are losing their income, and a growing share of us ask a machine instead.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

12 · Power
Yes, without migration.
Photograph: Unsplash
Read the chapter12 · Power
Power in 2035 will be shaped by two forces that move at very different speeds. Demography is slow and almost certain. Geopolitics is fast and almost unknowable. This chapter leans on the first and is careful about the second.
Full edition
The rest of this chapter is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

Part three · Worlds in 2035
Four worlds, one household, nine years.
Photograph: Unsplash
Read the chapterPart three
The twelve open questions in part two are not independent. They tend to resolve together, because the same two things sit under most of them.
Full edition
The rest of this part is in the full edition.
All twelve chapters, the four worlds and every source, as a 167-page download you keep. £550 + VAT.

Part four · The calls
Twelve calls, marked in public every October.
Photograph: Unsplash
Read the chapterPart four
Here are the twelve questions people ask us most about the next decade, our answer to each, and the call that tests it. Every call names the body whose published figure will settle it, so that nobody, least of all us, gets to mark it generously.
| Question | Our answer | The call for 2035 | Confidence | Settled by |
|---|---|---|---|---|
| Will AI take our jobs? | No. Not enough to break the jobs market. | UK unemployment is below 5.5% in February to April 2035 | 60% | ONS labour market statistics |
| Will AI eat the power grid? | Yes. | The world's data centres use more than 1,000 TWh of electricity in 2035 | 70% | International Energy Agency |
| Will we need a universal basic income? | No. We will pay for a bigger benefits bill instead. | The UK pays no universal basic income, and working-age benefit spending is above 4% of GDP in 2034-35 | 70% | DWP benefit expenditure tables |
| What happens to the NHS? | It survives, but misses its own promise. | Fewer than 75% of patients waiting for planned treatment in England have waited under 18 weeks in March 2035 | 55% | NHS England referral-to-treatment statistics |
| Will renewables win? | Only just. | Renewables generate at least half the world's electricity | 60 to 65% | Ember Global Electricity Review |
| Will we stop buying so much stuff? | We already have, per head. | Retail sales volume per person in Great Britain is still below its 2021 peak | 65% | ONS retail sales and population estimates |
| Will Britain make more of what it buys? | No. | The UK's deficit on trade in goods is at least 6% of GDP | 75% | ONS balance of payments |
| Will we build the homes we need? | No. | The seven new towns proposed in 2026 have completed fewer than 30,000 homes | 70% | Official housing completions |
| Will British farming survive? | Yes. Fewer farms, same share of food. | The UK still produces at least 60% of the food it eats, by value | 65% | Defra, Agriculture in the United Kingdom |
| Is university still worth it? | For most, yes. Not for all. | At least 35% of UK 18-year-olds go to university | 70% | UCAS end-of-cycle data |
| Will we get our news from AI? | Yes, nearly a third of us. | At least 30% use AI assistants for news weekly, and trust in news is between 30% and 40% | 60% | Reuters Institute Digital News Report |
| Is Britain shrinking? | Yes, without migration. | The UK records more deaths than births in each year to mid-2027 through mid-2035, and its population is under 72 million at mid-2035 | 65% | ONS mid-year estimates |
If these numbers mean anything, we should get about eight of the twelve right. If we get all twelve, we were too timid and should have said things that were harder to know. If we get five, we were overconfident and you should discount us accordingly. Being wrong on some of these is what honest confidence levels look like.
The odds in this section are synthetic: they come from language models asked to forecast, not from people, and they are published whole, including the ones that embarrass us. How the panel was built, and what it cannot do, is on the method page.
Before publishing we put the twelve calls to a panel of 24 synthetic forecasters: twelve analytical lenses (a labour economist, a demographer, an energy systems analyst, a public finance specialist, an infrastructure delivery specialist, a political scientist, an actuary and others), each run on two different models, each working alone, each told exactly how a call would be marked but not how confident we were, and each allowed to look up the latest official figures before answering. Ensembles of language models have matched human forecasting crowds in published tests, which makes this the one use of synthetic voices with evidence behind it. The panel's job is to tell us where we may be fooling ourselves, not to replace our judgement.
| Question | Our confidence | Panel median | Middle half of the panel |
|---|---|---|---|
| Will AI take our jobs? | 60% | 78%* | 69 to 85* |
| Will AI eat the power grid? | 70% | 76% | 74 to 82 |
| Will we need a universal basic income? | 70% | 64% | 63 to 70 |
| What happens to the NHS? | 55% | 87%* | 84 to 91* |
| Will renewables win? | 60 to 65% | 54% | 48 to 60 |
| Will we stop buying so much stuff? | 65% | 70% | 66 to 78 |
| Will Britain make more of what it buys? | 75% | 84% | 81 to 88 |
| Will we build the homes we need? | 70% | 84% | 70 to 92 |
| Will British farming survive? | 65% | 47% | 47 to 52 |
| Is university still worth it? | 70% | 55% | 50 to 57 |
| Will we get our news from AI? | 60% | 34% | 31 to 38 |
| Is Britain shrinking? | 65% | 57% | 53 to 76 |
* The panel judged the earlier, easier versions of these two calls (unemployment below 6%; fewer than 92% seen within 18 weeks). We made both harder after reading its view, so its figures for them are not comparable with ours.
The splits run both ways. On news from AI, farming and university the panel thinks we are too sure; on the NHS, trade and the new towns it thinks we are not sure enough.
Six of the calls were put to the panel twice. The first round, on 11 October, let each forecaster search for the latest figures, and it showed that retrieval mattered more than reasoning: the forecasters who missed a twelve-day-old announcement on the state pension sat 22 points below those who found it. So when we replaced six calls with plainer questions (on a basic income, the NHS, consumption, trade, farming and university), the second round gave every forecaster the same dated evidence pack and no search. Disagreements in that round are about judgement, not news.
Taken together, the panel expected about 7.9 of the twelve calls to come right; our own numbers imply 7.9 now, and 8.1 before the two changes described below. It is surer than we are of the calls where a long trend has to break for us to be wrong: the NHS missing 18 weeks, the trade deficit, the new towns, unemployment. It is less sure of the calls that turn on behaviour or a knife-edge threshold: news from AI most of all, then farming, university, renewables and the population. Its doubt on news is specific: it expects weekly use of AI for news to plateau in the twenties, as podcasts did, and its own test is the 2027 Digital News Report, where 14% or more keeps the call alive and 11% or less means it is stalling. We will apply that test in our first scorecard.
Where the panel gives us pause: farming and university. On farming it puts the call at 47%, a coin toss, because the UK's food self-sufficiency already sits close to our 60% line and has drifted down for decades. On university it puts us at 55%, because the entry rate for 18-year-olds slipped in 2025 and again on early 2026 figures and the panel weighs the falling graduate premium and the strain on university finances more heavily than we do. We have held both, and say so: if the 2027 figures move against us we will lower our confidence in public.
Where we listened: the NHS and jobs. The panel put our original NHS call, that the 92% standard would still be missed, at 87%. That told us the call was too easy, not that we were too cautious, so we made it harder: the call is now that fewer than 75% will have been seen within 18 weeks in March 2035, which bets that the recovery stalls well short of the promise rather than merely missing it. On jobs the panel put us at 78%, so we moved the line from 6% to 5.5%, below the level in the Office for Budget Responsibility's own scenario where technology pushes structural unemployment up. Our confidence in each is lower as a result, 55% and 60%. If the report's rule is that getting all twelve right means we were too timid, these are the two places we have tried to be less so.
The panel also found a marking hazard in every one of the twelve calls, from the ONS's switch to a new labour force survey to the absence of any official count of homes completed in new towns. The marking rules we will hold ourselves to are on the method page, because a call that cannot be marked is not a call.
Read down the answers and a view of the decade appears. Most of them say that something people fear or hope for does not arrive on time: AI does not empty the job market, the homes are not built, the NHS does not get back to 18 weeks, Britain does not make more of what it buys, the basic income does not come. Several say that a slow trend keeps going: more deaths than births, less stuff bought per head, a farming sector that shrinks without losing its share of the plate. Only a few say that something moves fast: the machines' appetite for electricity, the share of it that comes from renewables, and the number of people who get their news from a machine.
That is the argument of this report in twelve lines. Intelligence is getting cheap fast. Almost everything it needs in order to matter (people, proof, physical capacity and the ability to join them up) is scarce and slow.
We have not predicted the outcome of any war or election, and we have not put a date on any breakthrough in artificial intelligence. Those are the things people most want forecast and the things least possible to forecast. We would sooner say so.
Every October The Future Collective will publish where each call stands: on track, off track or settled. Where the evidence has moved, we will say how our confidence has changed and why. If we abandon a call we will say that too, and leave the original on the page.
A forecast tells you what to expect. A question tells you what to do. These are the twelve we would put to any leadership team, and to ourselves.
Method
We will build a synthetic population of UK adults and use it in three ways: to test it against today's Britain, to live out the four worlds to 2035, and to put independent odds on the twelve calls. Every synthetic number will be labelled as synthetic, and none will be presented as a measurement of real people.
Three layers, each with a different job.
| We will say | We will not say |
|---|---|
| "In our simulation, households like X tend to do Y under world Z" | "N% of Britons will do Y by 2035" |
| "The synthetic population reproduces today's figure for X to within N points" | "Our model predicts" |
| "Real people in our interviews told us" (kept separate, always) | Any synthetic figure without its label |
Forty-eight archetypes span age (four bands), region (London; the South; the Midlands and North; Scotland, Wales and Northern Ireland) and household income (three bands), with tenure, work status, household type, health, ethnicity and 2024 vote assigned to match published totals. Each archetype was run twice on Claude Sonnet and once on Claude Fable 5.1, 144 synthetic responses in all. Each run invents a specific person within the archetype's fixed attributes, describes an ordinary Tuesday in their voice, and then answers the battery below as that person would answer a market researcher in October 2026. Option order and question order are shuffled per run, each run is a fresh context, and the personas are not steered towards the right answer; the point is to measure the raw bias.
| Question | Real figure | Source |
|---|---|---|
| Works a hybrid pattern (among workers) | 28% of workers in Great Britain; 45% of those earning £50,000 or more, 8% of those under £20,000; 41% of degree holders; 19% of workers aged 16 to 29, 36% of 30 to 49s, 24% of 50 to 69s | ONS, June 2025 |
| Paid for private healthcare in the last year | 16% of adults in England | Healthwatch England, via trade press |
| Holds at least three-quarters of savings in cash (of those with over £10,000) | 61% | FCA Financial Lives 2024 |
| Took regulated financial advice in the last year | 8.6% of adults | FCA Financial Lives 2024 |
| Often or always lonely | 6.6% of adults in England; 9% of 16 to 24s | Community Life Survey 2024/25 |
| Trusts the UK government (6 or more on a 0 to 10 scale) | 27%; the courts and legal system 62% | ONS, fieldwork October 2023 (the OECD's 2025 wave was published in June 2026; UK figure to be added) |
| Household ran short of food in the last month | 12.2% of households; 16.7% of those with children | Food Foundation, June 2026 |
| Pays for online news | 10% in the UK | Reuters Institute, 2026, UK page |
| Uses an AI chatbot for news weekly | 4% in the UK, the lowest of 48 markets and flat on the year; 10% across all markets; 16% of under-35s | Reuters Institute, 2026, AI chapter |
| Trusts most news most of the time | 30% in the UK, down five points in a year | Reuters Institute, 2026, UK page |
| Uses generative AI for any purpose | 61% have ever used one and 34% use one weekly (six-country average including the UK, mid-2025); 40% of the UK public had used an LLM by late 2024 | Reuters Institute, Generative AI and News 2025; Alan Turing Institute and Ada Lovelace Institute, 2025 |
| Home has a heat pump | Under 2% of UK homes | Climate Change Committee, 2026 |
| Smokes cigarettes | 9.8% of UK adults; 7.2% of 18 to 24s, 12.4% of 45 to 54s, 7.2% of over-65s | ONS, Adult smoking habits in the UK: 2025 |
| Vapes daily or occasionally | 9.9% of adults in Great Britain | ONS, 2025 |
| Expects children today to be better off than their parents | 19% better off, 79% worse off | Pew Research Center, spring 2024 |
| Expects own finances to worsen in the next year; uses social media for news weekly | Asked for information; no anchor used in the pilot | ONS and Reuters track both |
Two further checks test the population weights rather than the personas: work status against the ONS labour market figures (employment 75.1%, unemployment 4.9%, inactivity 20.9%), and tenure against the English Housing Survey.
The research on synthetic respondents is young, mostly American, and largely sceptical. The main findings, each of which shaped the pilot design and each of which the pilot then reproduced.
Right averages, wrong distributions. Bisbee and colleagues (Political Analysis, 2024) had ChatGPT answer American National Election Study questions as thousands of described respondents. The averages matched the real survey closely; the variation was smaller than in real responses, regression coefficients often differed significantly, the distributions moved with minor changes of prompt wording, and the same prompt gave different results three months apart.
A bias of the model's own. Boelaert and colleagues (Sociological Methods and Research, 2025) call it machine bias: on each topic a model shows a strong bias and low variance, and the bias changes unpredictably from one topic to the next. Their conclusion is that current models cannot yet replace research subjects in opinion or attitude research.
Caricature. Wang, Morgenstern and Dickerson (Nature Machine Intelligence, 2025) show that models asked to speak as a group produce portrayals closer to how outsiders stereotype that group than to how its members describe themselves, and flatten the variety inside the group.
Survey mechanics. Dominguez-Olmedo, Hardt and Mendler-Dünner (NeurIPS 2024) found a strong preference for whichever option is listed first and, once that is corrected, answers that resemble no real demographic group. Synthetic respondents almost never say they don't know.
The largest independent test, published on 30 September 2026. Pew Research Center (30 September 2026) built digital twins of about 6,700 members of its American Trends Panel and re-ran three 2026 surveys, nearly 300 questions, through Claude Opus 4.6 and GPT-5.1. The synthetic results differed from the real ones by an average of 12 percentage points (11.4 for Opus, 13.3 for GPT), by more than 15 points on 28% of questions, and by most for Republicans (16.1 points) and Black adults (15.1). Nearly half the questions had at least one answer option that no synthetic respondent chose. The twins knew far more than real people (98% could name a First Amendment right, against 52% of humans) and said "not sure" a quarter as often. The two models erred in opposite directions: GPT made Americans more extreme than they are, Opus more middle-of-the-road. Pew's verdict: not an adequate replacement for polling.
The professional standard. The AAPOR task force on AI in survey research (2026) calls synthetic responses the riskiest use of AI in the field, says there is no consensus on when, if ever, they can stand in for human ones, and draws the acceptable boundary at clearly labelled pretesting, pilot work and exploratory diagnostics. It prefers the term synthetic responses to synthetic samples, because the method is not a sampling design, and says the next tests should ask whether correlations, joint distributions and subgroup interactions are preserved, not only the marginal totals.
Where the evidence is better. Two findings cut the other way, and both are built into the plan. Park and colleagues (Stanford, 2024) built agents for 1,052 real people from two-hour interviews; those agents reproduced each person's own General Social Survey answers at 83 to 86% of the person's own two-week test-retest consistency, against 74% for agents built from demographics alone, with smaller accuracy gaps across racial and ideological groups. And Schoenegger and colleagues (Science Advances, 2024) found that an ensemble of twelve language models matched a human forecasting crowd on a set of real-world questions, which is the evidence behind using a synthetic panel to put odds on the twelve calls.
What this meant for the pilot. Option order and question order shuffled per run; one respondent per fresh context; a specific invented person rather than a demographic label; 2024 vote included in every persona, because politics predicts attitudes; raw, unsteered answers so the bias can be measured; subgroup and within-archetype checks, not just totals; and every figure labelled synthetic. One weakness no design choice can fix: an answer that no synthetic respondent ever gives cannot be reweighted into existence afterwards.
All figures in this section are synthetic responses, weighted to the UK adult population by archetype. They are not measurements of real people. The real figure sits beside each one.
The run. 11 October 2026. Every response kept the attributes it was given: 144 of 144 matched the assigned work status and the presence of children, so the population layer does its job of holding the structure fixed. The archetype weights reproduce the ONS shares for age, region, income, tenure and work status to within a point. Ranges in brackets are 90% intervals from resampling the 48 archetypes, which is the honest size of the pilot's uncertainty.
| Measure | Real | Sonnet (96 responses) | Fable (48 responses) |
|---|---|---|---|
| Hybrid working, among workers | 28% | 32% (17 to 48) | 32% (17 to 50) |
| Hybrid, workers earning £50k or more | 45% | 90% (79 to 100) | 81% (57 to 100) |
| Hybrid, workers earning under £20k | 8% | 0% | 0% |
| Hybrid, workers with a degree | 41% | 71% (50 to 92) | 71% (50 to 92) |
| Hybrid, workers aged 18 to 34 / 35 to 54 / 55 to 74 | 19% / 36% / 24% (ONS bands differ) | 29% / 38% / 25% | 29% / 39% / 25% |
| Works from home only, among workers | a stable minority | 0% | 0% |
| Paid for private healthcare, last 12 months | 16% (England) | 3% (1 to 6) | 25% (15 to 35) |
| Holds three-quarters or more of savings in cash, of those with over £10k | 61% | 65% (48 to 82) | 59% (42 to 75) |
| Took regulated financial advice, last 12 months | 8.6% | 0% | 14% (6 to 21) |
| Lonely often or always | 6.6%; 9% of young adults | 0% | 0% |
| Trusts the UK government, 6 or more out of 10 | 27% | 0% | 0% |
| Trusts the courts, 6 or more out of 10 | 62% | 34% (24 to 45) | 61% (50 to 73) |
| Household food insecure in the last month | 12.2%; 16.7% with children | 21% (11 to 30); 24% with children | 17% (8 to 27); 26% with children |
| Pays for online news | 10% | 6% (2 to 11) | 18% (8 to 27) |
| Used an AI chatbot for news last week | 4% (UK) | 0% | 0% |
| Used social media for news last week | not anchored | 75% (66 to 84) | 80% (71 to 90) |
| Trusts most news most of the time | 30% | 10% (4 to 16) | 38% (27 to 50) |
| Currently uses AI tools for any purpose | 34 to 61% | 59% (48 to 70) | 50% (40 to 62) |
| Uses AI tools, aged 18 to 34 / 55 to 74 / 75 and over | steep gradient, not vertical | 100% / 20% / 0% | 100% / 13% / 0% |
| Home heated by a heat pump | under 2% | 0% | 0% |
| Smokes cigarettes | 9.8% | 5% (2 to 8) | 4% (0 to 10) |
| Smokes, aged 18 to 34 | 7% and up | 0% | 0% |
| Vapes | 9.9% | 12% (5 to 20) | 21% (12 to 31) |
| Expects children today to be better off / worse off | 19% / 79% | 0% / 100% | 0% / 100% |
| Expects own finances to get worse / better in the next year | not anchored | 62% / 0% | 39% / 3% |
The headline averages come out close because the errors cancel: whole groups vanish at one end while the comfortable majority is overdrawn at the other.
Mean absolute error against the 27 anchors: Sonnet 11.6 points, Fable 9.6 points. Pew's independent test of the best available model on 300 American questions came out at 11.4. We are where the state of the art is, and the state of the art is not good enough to publish as measurement.
What the pilot shows
Our call on the method
As a measuring instrument, the raw method fails the calibration test, in exactly the ways the literature predicted: plausible totals, caricatured subgroups, missing minorities, and a tilt that depends on which model you ask. We will say so in the report, in full, and we will not publish a synthetic level as if it described Britain.
It is not useless, and the pilot says how to use it.
The pilot's own numbers, scripts, archetype definitions and all 144 synthetic responses are available from The Future Collective on request so any figure here can be rerun.
Synthetic. Twenty-four forecasts per call from language models, not people.
Design. Twelve analytical lenses (a base-rate generalist, a labour economist, an energy systems analyst, a public finance and pensions specialist, a demographer, a health inequalities epidemiologist, an infrastructure delivery specialist, a retail and consumer analyst, a media researcher, a historian of technology adoption, a political scientist and an actuary), each run once on Claude Fable 5.1 and once on Claude Sonnet, each run covering six of the twelve calls in shuffled order: 48 runs, 24 independent forecasts per call. Every forecaster was given the call, the settling body and the exact marking rule, and was not told our confidence. Each could use up to two web searches per call to find the current value of the indicator, and had to write down its reading of the rule, the current state, an outside view, the strongest case each way, the likeliest failure path, a probability from 1 to 99, its confidence in its own number and the one 2027 figure that would change its mind. Compound calls were estimated part by part.
| Call | Our confidence | Panel median | Middle half | Full range | Fable / Sonnet medians |
|---|---|---|---|---|---|
| Work | 65% | 78% | 69 to 85 | 68 to 87 | 69 / 85 |
| Machines | 70% | 76% | 74 to 82 | 59 to 86 | 73 / 82 |
| Money | 75% (was 60%) | 44% | 42 to 49 | 33 to 73 | 48 / 42 |
| Health | 75% | 82% | 78 to 86 | 62 to 91 | 82 / 82 |
| Energy and climate | 60 to 65% | 54% | 48 to 60 | 46 to 81 | 58 / 52 |
| Buying | 65% | 63% | 57 to 71 | 41 to 80 | 56 / 72 |
| Making and moving | 85% | 94% | 93 to 97 | 89 to 98 | 93 / 97 |
| Places | 70% | 84% | 70 to 92 | 61 to 95 | 69 / 92 |
| Food | 75% | 72% | 69 to 79 | 61 to 86 | 80 / 70 |
| Learning | 75% | 93% | 92 to 95 | 86 to 97 | 92 / 95 |
| Truth | 60% | 34% | 31 to 38 | 24 to 44 | 34 / 32 |
| Power | 65% | 57% | 53 to 76 | 14 to 88 | 54 / 77 |
What the panel taught us about the panel.
Marking rules. The panel found a hazard in every call. These are the rules we will mark ourselves by.
The panel's 48 runs, the assignments each forecaster received and the aggregation script are available from The Future Collective on request.
Synthetic. Twenty-four forecasts per call from language models, not people.
On the evening of 11 October we replaced six calls with plainer questions: a basic income (Money), the NHS (Health), consumption (Buying), trade (Making and moving), farming (Food) and university (Learning). The same 24 forecasters, twelve lenses on two models, judged the new calls, but this time with no web search: each was given the same dated evidence pack of 11 or 12 sourced facts per call, so that disagreement would be about judgement rather than news.
| Question | Our confidence | Panel median | Middle half | Fable / Sonnet medians |
|---|---|---|---|---|
| Will we need a universal basic income? | 70% | 64% | 63 to 70 | 62 / 70 |
| What happens to the NHS? | 70% | 87% | 84 to 91 | 84 / 91 |
| Will we stop buying so much stuff? | 65% | 70% | 66 to 78 | 67 / 79 |
| Will Britain make more of what it buys? | 75% | 84% | 81 to 88 | 80 / 88 |
| Will British farming survive? | 65% | 47% | 47 to 52 | 47 / 50 |
| Is university still worth it? | 70% | 55% | 50 to 57 | 57 / 50 |
After this run we made two calls harder: unemployment below 5.5% (was 6%) and fewer than 75% of NHS patients seen within 18 weeks (was 92%). The tables above record the confidence and panel figures for the earlier versions.
With a shared evidence pack the spread narrowed sharply: no call's middle half was wider than 13 points, against 24 in the first round. Sonnet again sat above Fable, on five of the six; university was the exception.
Marking rules for the new calls
What next for you
For you
£550+ VAT
The complete report, to download and keep.
For your leadership team
£3,500+ VAT · half day
We present the report to your team, then help you work out the so what: what it actually means for your company.
Or email matt@future-collective.co.uk
For your leadership team
£8,000+ VAT · full day
We take the parts of the report that matter most to your company and industry, and run a full-day workshop with your team.
Or email matt@future-collective.co.uk