No one can honestly promise a job is safe from AI. But “will AI take my job” is only useful up to a point: the evidence increasingly shows that exposure isn’t set by your industry or the title on your payslip. It depends far more on what you actually do all day, which of your tasks a machine can already do, and which still need you.

Why this question gets asked so often

The concern is recent and it is rising fast. Google Trends data shows search interest in “will AI take my job” and “jobs safe from AI” climbing sharply since 2023, in step with how quickly AI tools have spread through ordinary workplaces.

Often, though, the question isn’t really “should I change career” yet. It’s earlier than that. Most people typing this into a search bar are trying to work out whether there’s a threat at all, before they’ve got anywhere near deciding what to do about it. That’s a different job than a retraining decision, and worth noticing, because doom-scrolling AI headlines with no plan attached is exactly what turns a reasonable concern into constant, low-grade dread.

What the evidence actually shows

Three findings are worth knowing before assuming the worst.

The major labour-market projections do not currently point to an overall collapse in employment. The World Economic Forum’s Future of Jobs Report 2025 projects around 170 million roles created globally by 2030 and 92 million displaced, a net increase of 78 million. It’s worth being precise about what that figure actually measures: it reflects the combined effect of several forces reshaping work, including AI, demographic change, the green transition and economic pressure, not AI on its own. Within that total, the Forum’s own estimate for AI and information-processing technologies specifically is smaller and more balanced: around 11 million jobs created and 9 million displaced. The churn is real, and for the person whose specific role disappears, a positive global number is not much comfort. But it is very different from evidence that AI is simply removing work from the economy wholesale.

So far, the evidence on employment is more mixed, and more encouraging, than the headlines suggest. PwC’s 2026 Global AI Jobs Barometer, which analyses more than a billion job postings across six continents, found headcount growing faster at companies most exposed to AI than at those least exposed. Its 2025 edition found something similar at occupation level: job numbers in highly AI-exposed occupations grew 38% between 2019 and 2024, compared with 65% in less-exposed occupations. Slower, but still real growth, not decline. None of this proves AI caused the hiring; correlation isn’t causation here. But it does challenge the simple assumption that more AI exposure automatically means fewer jobs.

Risk tracks task type more than job title, and skill is being rewarded as much as it’s being threatened. Displacement clusters around routine, repetitive, clerical and administrative work, the parts of a job that are easiest to specify and hand to a machine. Roles that depend heavily on contextual judgement, accountability, complex decision-making and human interaction tend to be less exposed to what today’s AI systems can do. At the same time, PwC’s 2026 analysis found workers with AI skills commanding an average 62% wage premium across the countries it covers, up from 57% the year before. The same disruption creating risk on one side is creating real, well-paid demand on the other.

What is genuinely different depending on your role

It would be dishonest to pretend the risk is spread evenly, or to file this under “probably fine.” The World Economic Forum’s own figures suggest 41% of employers expect to reduce their workforce specifically where AI can automate tasks, alongside 77% who plan to upskill existing staff and roughly half who expect to redeploy people out of exposed roles into other parts of the business. Those three numbers sitting side by side are the honest picture: AI is already changing work materially, some roles will genuinely shrink, and most employers’ first response is to retrain and move people rather than simply let them go. Which of those you land closest to depends on what your actual work is made of, not on your job title.

The scale of change shouldn’t be understated either. The World Economic Forum estimates that job creation and displacement combined could amount to 22% of today’s formal jobs by 2030. That is a meaningful shift, and nobody should be reassured out of taking it seriously.

What none of this tells you is your own timeline. Nobody can name the date your specific role becomes exposed. What the evidence does support is that people building AI-adjacent, AI-complementary skills now are already being rewarded for it, ahead of any crisis point, not after one.

A more useful question than “will AI take my job”

“Is my job safe” rarely has a clean answer, because safety here isn’t a fixed property of a job title. It’s a moving read on task exposure, transferable skill, and how much runway you have to adapt. The more useful questions are structural: how much of what you do day to day is genuinely hard to automate, what would a plan B look like if your role narrowed rather than vanished outright, and whether retraining now, while you still have time and choice, beats retraining later, under pressure.

Those are exactly the questions Should I Retrain is built to help you work through. It’s a short, structured diagnostic that looks at your specific situation rather than industry-wide averages, and gives you an evidence-informed view of whether retraining makes sense for you now.

Take the Should I Retrain diagnostic →

Last updated: July 2026.

Sources: World Economic Forum, “Future of Jobs Report 2025”; PwC, “Global AI Jobs Barometer” (2025 and 2026 editions); International Labour Organization, “Generative AI and Jobs: A Refined Global Index of Occupational Exposure” (2025).