What Jobs Will AI Replace by 2030?
Office and administrative support, customer service, and entry-level software roles face the heaviest task-level exposure by 2030, according to the three most-cited forecasts on this question. None of them say those jobs disappear outright. They say a large share of the tasks inside those jobs get automated, which is a different claim, and the difference matters for anyone trying to plan around it.
This page names the actual forecasts, where they overlap, and where they contradict each other, so you're not relying on a single number that happens to be the scariest one you've seen shared online.
The three forecasts, named
The World Economic Forum's Future of Jobs Report 2025 surveyed more than 1,000 of the world's largest employers directly about their hiring plans. It projects 92 million jobs displaced globally by 2030 against 170 million created, a net gain of 78 million, with total churn touching about 22% of jobs worldwide (World Economic Forum). The same report found 41% of employers expect to reduce headcount specifically where AI can automate tasks, and that by 2027, nearly half of companies expect entry-level hiring to be eliminated at their organization, with 21% saying AI is the sole reason (World Economic Forum).
McKinsey Global Institute modeled automatable work hours in the US economy rather than headcount, and found generative AI could push the share of automated hours from about 21.5% to roughly 29.5% by 2030, an 8-point jump, with an estimated 12 million occupational transitions needed as a result (McKinsey). That research found the pressure isn't even: lower-wage workers face up to 14 times more pressure to change occupations than the highest earners.
Goldman Sachs took a task-exposure approach and put office and administrative support at 46% task exposure and legal work at 44%, against much lower exposure for physically demanding and outdoor occupations (CNBC; Goldman Sachs). Its 2023 analysis estimated 300 million full-time jobs globally exposed to some degree of automation, with up to a quarter of current US and European work potentially done entirely by AI.
Where the three agree
All three point to the same category of work: routine, digital-only, low-accountability tasks. Data entry, first-draft writing, basic customer support responses, scheduling, and standard document review show up as exposed in every version of this research. Microsoft Research's separate analysis of 200,000 real Copilot conversations found interpreters, writers, and customer service representatives scored highest on task overlap with what generative AI already does, while physical trades scored lowest (Microsoft Research).
Entry-level versions of these roles face more pressure than senior versions of the same job. Stanford economist Erik Brynjolfsson and colleagues analyzed ADP payroll data covering 25 million US workers and found a 13% relative decline in employment for 22-to-25-year-olds in the most AI-exposed occupations since generative AI adoption accelerated, while older workers in the same jobs held steady or grew (Fortune). That's a real, measured effect, not a 2030 projection, and it's concentrated in exactly the tasks juniors traditionally use to build skill.
Where the three disagree
WEF's framing is net-positive: more jobs created than destroyed. Goldman Sachs' framing leans more alarmist on raw exposure numbers. Microsoft's own researchers pushed back on reading high task-overlap as job elimination, warning explicitly against that interpretation of their own data (Microsoft Research blog). None of the three measure the same thing: WEF surveys hiring intent, McKinsey models hours, Goldman Sachs models task exposure. Treat them as three angles on the same rough pattern, not three votes on a single number.
PwC's 2025 Global AI Jobs Barometer adds a complicating data point that cuts against the doom framing: jobs with high AI exposure actually grew 3.5 times faster than other occupations in recent years, and workers with AI skills now command a 56% wage premium, up from 25% the year before (PwC). Exposure to AI and job growth are not opposites in the data PwC collected. That doesn't cancel out the McKinsey and Stanford findings on entry-level pressure. It means the 2030 story has both things happening in the same labor market at once.
One occupation where the number is already flat
Paralegals are a useful test case because BLS has already priced AI into the forecast rather than leaving it as a future risk. BLS projects paralegal and legal assistant employment to grow just 0.2% from 2024 to 2034, essentially flat, adding only about 600 net positions nationwide even as roughly 39,300 openings a year still occur from turnover and retirements (BLS). BLS names the reason directly: growth is limited by advances in technology, including AI, the same document-review and first-draft tasks Goldman Sachs flagged at 44% exposure for legal work overall. A flat national number, sitting alongside tens of thousands of annual openings from people leaving the field, is what task-level automation actually looks like in the data, not a headline collapse, but a ceiling on net new hiring.
What this means if you're planning around it
If your job involves a lot of routine drafting, basic support tickets, or data entry with little discretion attached, the task-level pressure described above is already showing up in the data, not just in the forecasts. If your job involves judgment calls, physical presence, client relationships, or work that requires someone accountable to sign off, the same research consistently shows lower exposure.
The honest version of a 2030 plan isn't "wait and see." It's build the parts of your role that require judgment now, while task automation is still handling the routine slice rather than the whole job.
Keep the skills that keep you employed
The tasks you can still do without leaning on AI are what make you hard to replace here. The free 5-Day AI Reset is a five-email course built around exactly that: Day 2 has you take one task back and do it unassisted. One small change per day, and it stays useful no matter which way by 2030 moves.
Frequently asked questions
What jobs are most at risk from AI by 2030?
Office and administrative support (46% task exposure per Goldman Sachs) and legal work (44%) rank highest, alongside customer service and entry-level software roles (CNBC). Physically demanding and outdoor occupations rank lowest.
Will AI create more jobs than it destroys by 2030?
The World Economic Forum's Future of Jobs Report 2025 projects a net gain: 170 million jobs created against 92 million displaced globally, about 78 million more jobs overall (World Economic Forum). That's a net figure across a churning labor market, not a guarantee for any individual role.
Are entry-level jobs disappearing faster than senior roles?
Yes, per real payroll data. Stanford's analysis of ADP records found a 13% relative employment decline for 22-to-25-year-olds in AI-exposed occupations, while older workers in the same jobs held steady (Fortune).
Does high AI exposure mean a job is shrinking?
Not necessarily. PwC found AI-exposed occupations grew 3.5 times faster than other jobs in recent years, alongside a 56% wage premium for workers with AI skills (PwC). Exposure and growth have coexisted so far.
Is paralegal work a good example of 2030-level exposure?
Yes. BLS projects paralegal employment to grow just 0.2% through 2034, essentially flat, and names AI directly as a limiting factor, even though the field still posts about 39,300 openings a year from turnover (BLS). That is task automation capping net growth, not a specific prediction of mass layoffs.
How is this different from the "next 10 years" or "2040" forecasts on this site?
This page focuses specifically on 2030, the horizon where WEF, McKinsey, and Goldman Sachs all still roughly agree on direction. See What Jobs Will AI Replace in the Next 10 Years? for a confidence-tiered view, and our 2040 outlook, publishing soon, for why longer horizons get much less reliable.
Keep reading
For a sector view, see What Jobs Are Safe From AI?.
See where you actually stand
Aggregate forecasts describe millions of jobs at once. Take the free How AI-Proof Is Your Job? assessment to score your own tasks instead of borrowing a global average.