🔍 Read the full analysis: Could AI Change Our Sense Of Purpose At Work? on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A 2026 survey of 8,234 working-age adults in the UK, US, Denmark and India found that most respondents would leave work for a guaranteed income below their current salary, but about one in five said they would stay even at 120%. The arXiv preprint argues that work can provide purpose, belonging and recognition that money may not replace; its findings are based on stated preferences, not observed behavior, and do not establish that AI causes changes in work or politics.
A new survey of 8,234 working-age adults finds that most would leave their jobs for a guaranteed income below their current salary, while about one in five would stay even at 120% of pay. The findings, in an October 29 arXiv preprint by Julian Jacobs of the University of Oxford and Google DeepMind, examine whether money can replace the purpose, social ties and recognition people say they get from work as AI reshapes employment.
The survey presented respondents with a hypothetical offer of guaranteed income, beginning at 60% of their salary and rising in 10-point steps to 120%. Healthcare, pension and paid leave were kept in the scenario. By an offer equal to 80% of salary, a slight majority said they would quit; at 100%, about three in four said they would leave. Among those willing to accept an offer, the median quit point was 70% of salary.
Still, about 20% said they would not leave at 120%. The paper reports similar holdout shares across income groups: 21% among lower-income respondents, 18% in the middle-income group and 22% among higher-income respondents. The figure was 21% among white-collar workers and 17% among blue-collar workers. These are answers to a hypothetical question, not records of what respondents later did.
For respondents the researchers could classify by their leading work value, 33% prioritized the work itself, including purpose, learning or contribution; 23% prioritized colleagues and belonging; 17% status and recognition; and 26% pay and benefits. The paper says 73.7% placed a nonfinancial reason first. In adjusted analyses, purpose-driven respondents asked for about 1.9 to 5.8 percentage points more guaranteed income before saying they would quit than pay-driven respondents, depending on the method used. The paper is a 39-page, single-author preprint and has not been peer-reviewed.
WHAT WORK IS FOR, ACCORDING TO WORKERS
Three in four put something other than pay first
The paper sorts workers by what they rate most important in their job. In its own figure, 73.7% of classified workers put a non-money reason first.
The work itself: purpose, learning, contributing to society.
Pay and benefits come first.
Belonging, colleagues, meeting people.
Recognition and prestige.
Pay-first share, by country
Share of classified workers whose top value is pay and benefits
Meaning is not a luxury
Share driven by the work itself (intrinsic), by household income third
Also about a third in craft and trade jobs (33%) and in services and sales (31%).
Shares of the 6,224 working respondents the paper could classify; 2,010 respondents (non-working, or no clear top value) were left out. 2,000 tied answers were assigned by a tie-break rule. Satisfied with their job: 81%.
Source: Jacobs, What is Work For?, arXiv:2610.39375v1, 29 Oct 2026 · preprint, not peer-reviewed · survey of 8,234 working-age adults in the UK, US, Denmark and India by More in Common, June–July 2026, supported by Google DeepMind.
Powered by Thorsten Meyer AI
What Income Cannot Measure
The findings matter to employers and policymakers weighing how automation could affect workers beyond wages. If jobs provide purpose, belonging or recognition, replacing lost earnings alone may not reproduce everything people value in their working lives. Jacobs argues that a basic income may be needed during an AI-related labor shock but could fail to compensate for these nonfinancial benefits. That is the author’s interpretation of survey responses, not a measured outcome of a basic-income program.
The results also caution against treating workers as a single group. Most respondents in the survey said they would leave under the hypothetical income offer, while a substantial minority preferred to keep working even when offered more than their current salary. The paper finds that purpose-driven respondents required somewhat more income to quit than pay-driven respondents, but the estimated differences were small in practical terms: at a $60,000 salary, the source estimates roughly $1,100 to $3,500 annually.
For organizations introducing AI, the study suggests a question worth testing locally: which parts of a job do employees value, and which could automation remove or change? The survey does not show that every job supplies meaning, or that redesigning work will preserve it. It does indicate that a discussion focused only on pay and job counts would miss reasons some respondents gave for staying.
Products Worth Considering
Samsung Essential S3 27-Inch Curved FHD Monitor
HP Sprocket Photo Booth Machine Instant Color Photo Printer with 10.1 Touchscreen, LED Ring Light, 250+ Frames, 3×4 ZINK Paper, QR Code Shar
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
As an affiliate, we earn on qualifying purchases.
How the Survey Tested Work
Jacobs’s paper, What is Work For? Labor, Politics, and Meaning in the Age of AI, was posted to arXiv on October 29, 2026, in the General Economics category. It draws on an online survey fielded in June and July by polling organization More in Common. Participants included 2,032 people in Denmark, 2,072 in India, 2,052 in the UK and 2,078 in the US. The survey used national quotas for age, gender, education and region.
Alongside the income-offer exercise, participants answered questions about job values, how they would spend time if they stopped working, perceived AI threats and political views. The analysis uses cross-sectional regressions controlling for factors including income, occupation, age, gender, education, part-time work and country. The paper says the survey was supported by Google DeepMind, Jacobs’s employer, and notes that its views do not necessarily reflect those of Oxford or DeepMind. The author is affiliated with both institutions.
The AI-related responses describe expectations, not verified job losses. In the survey, 40% of white-collar workers said AI was at least somewhat likely to replace their job within ten years, compared with 29% of manual and service workers. Those figures capture respondents’ assessments at the time of the survey; they do not predict how many jobs AI will actually eliminate.
“A basic income “may indeed be necessary” in an AI labor shock, but “could ultimately fail to compensate workers” for what work gives beyond money.”
— Julian Jacobs, author of the arXiv preprint
Limits of the Survey Evidence
The results show what respondents said they would do in a hypothetical scenario. They do not establish whether people would actually leave a job at the stated income levels, whether work supplies the same benefits across occupations, or whether guaranteed income could replace those benefits in practice. The survey is cross-sectional, so it cannot establish that AI concerns cause changes in workers’ attitudes or political views.
The paper reports that purpose- and community-oriented workers become more pessimistic when they expect AI to harm jobs in their industry. However, its tables do not show significant differences between the work-value groups in populist vote intention, and Jacobs describes the analysis as relational rather than causal. The paper is also not peer-reviewed; the author says its numbers could change in later versions.
Country comparisons need care. For example, the reported 28.8% holdout share in India came from an online sample that the paper describes as younger, more urban and more educated than the country as a whole. The survey therefore does not establish that this figure represents all working-age adults in India. The supplied findings do not settle how responses vary across all occupations, or how workers would respond to actual AI-driven changes in their jobs.
Peer Review and Workplace Evidence
The immediate next step is further scrutiny of the preprint’s methods and findings. Because the paper is an early version, later revisions or peer review may alter its estimates or interpretation. The source does not state a publication date for a journal version or announce a follow-up survey.
For now, the evidence supports a limited conclusion: many respondents would trade work for a secure income, while a minority say they would remain even at a higher guaranteed salary. Whether AI changes those choices will depend on how jobs and workplaces evolve. Actual employment outcomes, longitudinal research and evidence from workplaces adopting AI would be needed to test whether the stated preferences predict real behavior.
Key Questions
What did the survey ask workers?
It asked whether respondents would voluntarily quit if offered guaranteed income, starting at 60% of their salary and rising to 120%, while keeping healthcare, pension and paid leave in the hypothetical offer.
How many respondents said they would stay at 120% of salary?
About one in five said they would not quit even at a guaranteed income equal to 120% of their salary. The result is a stated response to a hypothetical scenario, not observed behavior.
Does the paper prove AI will cause people to lose purpose at work?
No. The survey asked about expected AI effects and work values, but its cross-sectional analysis does not show that AI causes changes in purpose, employment or political attitudes.
Has the study been peer-reviewed?
No. It was posted to arXiv on October 29, 2026, as a preprint and had not been peer-reviewed according to the source material. The author notes that figures could change in later versions.
Who funded the survey?
The paper says the survey was supported by Google DeepMind. Jacobs is affiliated with Google DeepMind and the University of Oxford; the paper says its views do not necessarily reflect those of either institution.
Source: ThorstenMeyerAI.com
