Does using AI make you less worried about your job?

Using AI more does not make people less worried about losing their job. In the largest dataset that asks, Gallup's panel of nearly 30,000 US worker observations published on 9 September 2026, workers who used AI daily or several times a week were more than twice as likely as occasional users to say their job was very likely to be eliminated within five years. The gap held when Gallup followed the same people over time. What went with a smaller gap was a feeling of being respected and cared for at work, and among frequent users those two things mattered far more than how much they used the tool. The people closest to the tool are the ones best placed to see what it is doing to their tasks, and more time with it leaves that fear where it was.

Twice the fear among the heaviest users#

Christos Makridis, writing for Gallup, pooled four waves of the Gallup Panel from 2023 to the first quarter of 2026. The panel recruits by probability sampling, so this is a weighted picture of employed US adults rather than a sample of people who chose to answer a survey about AI. About 30 per cent of respondents appear in two or more waves, which lets the analysis compare a worker with their own earlier answers.

The headline is plain. Frequent users, meaning daily or several times a week, were more than twice as likely as those who used AI a few times a month or a year to say their job was very likely to be eliminated within five years. In the first quarter of 2026 the figures were 6.3 per cent against 3.09 per cent. The looser measure, somewhat or very likely, took in roughly 19 per cent of all workers. The gap survived comparisons within the same broad occupation, demographic controls, and the within-worker comparison, which removes stable traits such as personality and baseline optimism.

The fear also costs something. Workers who expected elimination scored lower on engagement and job satisfaction, higher on burnout, and were more likely to be actively looking for another job.

The fear follows real exposure#

It would be convenient to read this as anxiety that more experience will cure. Gallup's own data points the other way: the higher the share of an occupation's tasks that generative AI can do, the more likely a worker in it is to fear being displaced. Frequent users are, on average, looking at a real risk. The same analysis adds a qualification that matters for anybody managing people: two workers with similar exposure can still report quite different levels of concern, which Gallup reads as organisational context shaping how the same objective risk is interpreted.

Where the gap narrows: respect and care#

Gallup measured five aspects of workplace quality, and two stood out. Workers who strongly agreed that they were treated with respect, or that the organisation cared about their wellbeing, were about six to seven points less likely to report displacement concern. For frequent users the association between heavy use and fear was 6.8 points smaller among those giving the highest respect rating, and 11.1 points smaller among those who felt cared for, set against the roughly 19 per cent who say elimination is somewhat or very likely. Both gaps held when the same worker was tracked over time. Among occasional users, strong or weak management made little difference.

Read that last finding carefully. Good management did not matter much to the people who barely used the tool, and mattered a great deal to the people who used it most. The heavier the use, the more the relationship with the employer carries.

A second Gallup reading, on a different question#

Six days later Gallup published its annual Work and Education poll, a separate telephone survey fielded from 3 to 24 August 2026. Lydia Saad reported that 27 per cent of US workers now worry technology could make their job obsolete, a new high, up from 20 per cent in 2025 and 13 per cent when the question was first asked in 2017. Among workers aged 18 to 44 the figure was 34 per cent, against 19 per cent for those 45 and over. The direction agrees with the panel. The two numbers cannot be added together or compared point for point: different mode, different wording, and an employed sample of 507 with a margin of plus or minus six points. This poll carries no breakdown by how often people use AI.

Five limits on reading this as cause#

• Observational, throughout. Nobody was assigned to use AI more or to a better manager. People already in threatened roles may use the tools more, and an organisation that treats people with respect may differ in other ways too.
• The strongest comparison is the least precise. Gallup says so itself: the within-worker estimates rest on the smaller group observed in several waves.
• Small base rates. "More than twice as likely" describes 6.3 against 3.09 per cent on the strict measure. A doubling of a small share is still a small share.
• The question is broader than AI. The item asks about new technology, automation, robots or artificial intelligence together.
• Expectations and one country. The panel measures what US workers expect, and in Gallup's words not jobs actually lost. Nothing here speaks to any other labour market.

One further line needs separating out. Gallup's article advises managers to explain the purpose of AI adoption, connect it to team capability rather than headcount efficiency, and create space for concerns. That is sensible advice and it is Gallup's recommendation; the measured variables were respect and felt care, and the advice was not itself tested.

Familiarity was supposed to be the cure#

This section is interpretation, kept apart from the evidence above.

A great deal of corporate AI adoption rests on an unstated theory of fear: people worry because they do not understand the tools, so get them using the tools and the worry fades. Usage dashboards are then read as a proxy for comfort. The Gallup panel puts that premise to a large, weighted sample, and it comes back the wrong way round. The people who use AI most are the people most convinced it will take their job.

Rahim Hirji described the mechanism from the inside in The Great Unbundling of Work on 25 May 2025, the essay whose line "Your Job Isn't Disappearing, It's Dissolving" has travelled furthest. Its account of his own reaction fits the panel closely: "That unease I used to have wasn't irrational. It was a warning." A frequent user watches their role being shredded into tasks, one task at a time, and each task the tool absorbs is evidence they can see and an occasional user cannot. On this reading the fear is information, and training people to use the tool more hands them more of it.

What the tool cannot supply is an answer to the question that information raises, which is what happens to me when those tasks go. That answer comes from the organisation, and the panel's management finding fits it: respect and care matter most exactly where use is heaviest. An adoption programme that counts logins and leaves that question unanswered is producing its most anxious employees among its most engaged ones, and then measuring the engagement as success.

What to say to the people using it most#

• Stop treating usage as reassurance. A rising adoption figure says nothing about how people feel, and on this evidence may go with more fear.
• Answer the task question, per role. Say which tasks are expected to move to the tool, what the person's time goes to instead, and how the role is expected to develop. Hirji's own framing is that jobs dissolve into tasks; people deserve to see the new shape as well as the old one breaking up.
• Put the frequent users first. The management effect in the panel sits almost entirely with them, and they are also the people an organisation most needs to keep.
• If you are the worried one, treat the worry as a prompt to map your exposure honestly, task by task, and move towards the parts that stay hard. Will AI replace my job sets out how.

Key sources

• Makridis, C. (2026). Using AI More Does Not Reassure Workers, Managers Do. Gallup, 9 September 2026. Gallup Panel, four waves 2023 to Q1 2026, nearly 30,000 worker observations. Graded entry.
• Saad, L. (2026). More U.S. Workers Fear Losing Their Jobs to Technology. Gallup, 15 September 2026. Telephone poll, 3 to 24 August 2026, 507 employed adults. Graded entry.
• Hirji, R. (2025). The Great Unbundling of Work. Box of Amazing, 25 May 2025. Argument, cited as interpretation rather than evidence.

Related SuperSkills research#

On assessing your own exposure, will AI replace my job. On which work holds up, which jobs are safest from AI. On why counting usage misleads, how to measure AI adoption properly. On where the time goes once the tasks move, what should happen to the time AI saves. On the professional identity bound up in the tasks, staying valuable in the age of AI.

About this research#

Rahim Hirji is the author of SuperSkills (Kogan Page, 2026), keynote speaker on AI and human capability, and founder of The SuperSkills Intelligence Company. Every figure on this page was read at source on 28 September 2026, including Gallup's own survey-methods notes, and both studies are graded in the evidence base. Displacement fear is Gallup's measure and not a SuperSkills term; the reading of it in the interpretation section is his.

How this research works  ·  Reviewed quarterly  ·  Found an error? Tell me and it is corrected on the page.