Which Jobs May Change Most in the Next Decade — and Why

AI will play a large role in which jobs see major gains and losses over the next decade, according to new Labor Department data. But the better frame, labor researchers argue, is not which jobs vanish — it is which tasks change, which grow, and which get redesigned. The stakes are concrete: the US economy could create more jobs than automation replaces, yet the shift may require the largest and most sustained workforce transformation in US history.

The real unit of change is the task, not the job:

Lists of jobs that will disappear make for sharp headlines, but they can hide how work actually changes. Most jobs combine routine steps with judgment, communication, physical action and responsibility. A tool may handle one part of that bundle while leaving the rest intact.

That is why the coming decade is better understood as a period of task change. Some roles will shrink, some will grow, and many will be redesigned. The outcome will vary by employer, region and regulation.

The practical implication is that a job title alone says little about risk. Two workers with the same title can face very different futures depending on which tasks dominate their day and how their employer chooses to adopt new tools. Region and regulation add further variation, since the same technology may be deployed faster in one market than another.

Which work is most exposed:

Work that follows predictable patterns is generally more exposed to automation or AI assistance. Examples discussed in labor research often include repetitive administrative processing, standardized content handling and routine service interactions. Exposure, however, is not the same as elimination.

Organizations still weigh accuracy, liability, customer trust, cost and integration. A task that can be automated in theory may remain human-led in practice, or it may become a human-plus-tool workflow. That distinction matters for workers, because a tool that assists a person leaves the judgment and accountability in human hands.

The gap between what is technically possible and what is actually deployed is where many predictions go wrong. A process that looks automatable on paper can stall on liability, on customer trust, or on the cost of integrating it into existing systems. In that gap, jobs persist — often reshaped rather than removed.

The human parts that slow the shift:

Roles built around trust, complex decision-making, care, skilled trades in unpredictable settings or direct accountability may change more slowly. Even in those fields, software can alter scheduling, documentation, diagnosis support or communication. The headline “disappear” rarely captures that mixed reality.

The point is not that these roles are immune. It is that the change shows up inside the job — in how it is documented, scheduled, supported — rather than in the job's disappearance. A nurse, a plumber, or a manager may use new software daily and still be the one accountable for the outcome.

Workers are not passive in this transition. Skills that travel across tools — clear communication, problem-solving, domain knowledge and the ability to supervise technology — can reduce the risk of being trapped in a single narrow task. Those skills are portable precisely because they are not tied to one piece of software or one employer's process.

What the projections say:

The projections behind these debates are large in scale and vary by source. Here is how the main figures line up.

1. New jobs created:

About 170 million new jobs will be created this decade, according to the World Economic Forum's Future of Jobs Report 2025. That figure describes global job creation, not a guarantee for any single worker or region.

2. Jobs displaced:

One summary of the Future of Jobs Report 2035 cites 92 million displaced. Read alongside the creation figure, it suggests a net gain in jobs globally, but with severe dislocation for the people and places on the losing side.

3. Skills in flux:

The same summary puts 39% of skills about to change. That is the quieter number, and arguably the more useful one for an individual worker: it measures how much of what people know today will need updating.

4. Where growth concentrates:

A majority of the fastest-growing roles for 2034 will be in the health care and information and technology sectors, according to Bureau of Labor Statistics data. The BLS 2025-35 projections also track the fastest growing occupations overall, those with the most new jobs, and the fastest growing roles requiring a bachelor's degree or more and some postsecondary education.

What workers can do now:

A responsible forecast avoids fake precision. No one can responsibly promise exactly which jobs will vanish by a certain year. The practical approach is to track where tasks are changing, support credible training paths and evaluate claims against evidence as the next decade unfolds.

That means treating bold predictions with skepticism and looking for the task-level detail underneath. A headline about a job disappearing is a starting question, not an answer. The better question is which tasks inside that job are actually changing, and whether the worker has a path to the ones that remain.

“Lists of jobs that will disappear make for sharp headlines, but they can hide how work actually changes.”

What the Record Shows:

The coming decade is better understood as a period of task change, not whole-job elimination. Work that follows predictable patterns is more exposed, yet exposure is not elimination — organizations still weigh accuracy, liability, customer trust, cost and integration. About 170 million new jobs may be created this decade while 92 million could be displaced, and 39% of skills may change, with the fastest-growing roles concentrated in health care and information technology.

The open question is whether training and policy can keep pace with a transformation that may be the largest and most sustained in US history.