Skip to main content
GARGLOO

AI

Will AI Take My Job?

Will AI take your job? The WEF, IMF, Goldman Sachs and US jobs data disagree on the numbers but agree on one thing: tasks change first.

Last updated: 2026-11-15 · 7 min read

The short answer: probably not your whole job, but parts of it

If you are worried that artificial intelligence will take your job, you are not alone. In November 2026, searches about an AI Job Impact Report jumped about 800% in a single day, and the question of whether AI will replace human work is now one of the most-asked career questions of the year. The honest answer from the research is less dramatic than the headlines, and more complicated.

No major study predicts that AI will simply delete a fixed share of all jobs. What the research does show is that AI is very good at taking over specific tasks inside jobs: drafting a first version of a document, summarising a long report, answering routine customer questions, or sorting data. A job is a bundle of tasks, and when some of those tasks are automated, the job changes shape rather than disappearing overnight.

That distinction matters, because it changes what you should actually pay attention to. The useful question is not only whether your job title survives, but how much of your working day is made of tasks that AI already does well, and how much of it depends on judgement, physical skill, trust, or dealing with people in person.

What the big numbers actually say

The most quoted forecast comes from the World Economic Forum. Its Future of Jobs Report 2025, based on a survey of more than 1,000 employers, projects that by 2030, 170 million new jobs will be created and 92 million displaced, a net gain of 78 million. The same report says the disruption will touch about 22% of today's jobs, and that employers expect 39% of workers' core skills to change by 2030.

Those numbers come with an important warning that often gets lost. The World Economic Forum counts disruption from technology, demographics, economic pressures and other forces together, not AI alone. And a net gain of 78 million jobs does not mean the same people keep working: the jobs created and the jobs displaced are usually different jobs, in different places, needing different skills.

The International Monetary Fund puts the exposure differently. Its research estimates that almost 40% of employment worldwide is exposed to AI, rising to about 60% in advanced economies, where more work involves cognitive and information-processing tasks. Exposure, in the IMF's language, means AI can affect the tasks in a job. It does not mean the job disappears.

Goldman Sachs gave the debate its most repeated number in March 2023, when its economists estimated that generative AI could expose the equivalent of 300 million full-time jobs worldwide to automation, with about 18% of work globally exposed. The same report stressed that most jobs are only partly exposed and therefore more likely to be complemented than replaced, and that about two-thirds of occupations in the United States and Europe are exposed to some degree of automation.

Tasks, not jobs: how AI actually enters a workplace

The US Bureau of Labor Statistics, which publishes the official ten-year job projections for the United States, makes the same point in plainer language: high exposure to AI does not necessarily mean employment will decline. Its projections build in AI effects only where it finds convincing evidence, and it describes the uncertainty around AI's impact as very high.

Two occupations show why the picture is mixed. Employment of customer service representatives is projected to fall 5.5% between 2024 and 2034, a decline of about 153,700 jobs, as simple tasks are increasingly handled by automated systems and self-service. Employment of software developers, by contrast, is projected to grow 15.8% over the same period, adding more than 267,000 jobs, even though AI coding assistants are now part of daily software work.

The pattern is consistent: AI takes the routine, repeatable parts first. A customer service role that is mostly answering the same five questions is far more exposed than one that handles complaints, negotiates, or solves unusual problems. A software developer who mostly writes boilerplate code faces more pressure than one who designs systems, talks to users, and makes trade-offs. The tasks decide the risk, not the job title alone.

Which kinds of work are most exposed, and which look safer

The Goldman Sachs analysis found exposure concentrated in office work: about 46% of tasks in administrative support roles and about 44% in legal work were judged exposed to automation, compared with single-digit shares in construction, installation and maintenance, and building and grounds work. Routine cognitive work, in other words, is where AI is strongest, because it runs on text, numbers and patterns.

Work that looks less exposed shares a few traits. It happens in the physical world, where a robot still struggles: plumbing, electrical work, cooking, cleaning, care work, and skilled trades. It depends on trust and human relationships: teaching young children, nursing, therapy, and management. Or it depends on judgement in situations that are new every time: handling a crisis, negotiating a deal, or making a decision where the rules are unclear.

None of this is a guarantee. Exposure estimates describe what AI can technically do, not what employers will actually do, and history shows that automation often changes demand in unexpected ways. Cheaper, faster work can create more demand for it, and entirely new roles appear that nobody forecast. The safest reading of the research is that routine, predictable, text-based work faces the most pressure, while physical, social and judgement-heavy work faces the least, for now.

What the research says you can actually do

The one finding every major study agrees on is that skills are the moving part. The World Economic Forum expects 39% of workers' core skills to change by 2030, and 85% of employers in its survey say they plan to prioritise upskilling. That is a large number, but it is also a practical one: the workers who adapt are usually the ones who learn to use AI tools for the routine parts of their work and spend their time on the parts AI cannot do.

A useful way to think about it is to split your own week into three piles: tasks AI already does well, tasks AI helps with but cannot finish, and tasks only a person can do. If the first pile is most of your week, that is a real signal, and the research suggests learning the tools, moving towards the second and third piles, or both. If most of your week is judgement, people, or physical work, the evidence says your job is changing more slowly.

No study, and no article, can predict what will happen to one individual reader's job. Forecasts describe whole economies, not single careers, and they have been wrong before. What the research can say, honestly, is this: AI is reshaping tasks faster than it is deleting job titles, the disruption is real and uneven, and the people who keep learning are the ones best placed whatever happens next.

Frequently asked questions

Will AI replace my job?

No study can answer that for one individual person. The research says AI replaces tasks before it replaces jobs: the World Economic Forum projects 92 million jobs displaced and 170 million created by 2030, while the IMF estimates about 40% of jobs worldwide are exposed to AI. Exposure means tasks are affected, not that the job disappears.

Which jobs are safest from AI?

The research points to work that is physical, social, or judgement-heavy: skilled trades, care work, teaching, and roles that depend on trust, negotiation, or handling new situations. Routine office and administrative work, where tasks are repetitive and text-based, shows the highest exposure in the Goldman Sachs and IMF analyses.

Will AI create new jobs?

The World Economic Forum's Future of Jobs Report 2025 projects 170 million new jobs created by 2030 against 92 million displaced, a net gain of 78 million. But the new jobs are usually different jobs needing different skills, so a net gain for the economy does not automatically mean an easy transition for every worker.

Should I change careers because of AI?

The research does not support panic career changes. A steadier response, supported by every major study, is to learn how AI tools work in your own field, build the skills AI cannot easily copy, such as judgement, communication and hands-on expertise, and keep watching how your own occupation is changing in official jobs data.

Sources

Will AI Take My Job? | GARGLOO