AI & the Future of Jobs
Explore jobs from highest to lowest AI exposure and see their projected outlook.
Most to least exposed to AI
↔ Swipe sideways to see all columns.
| Occupation | Job group | AI exposure | Job change | Projected timeline |
|---|
No matching occupations. Clear your search or broaden the filters.
The bigger picture
Key statistics & their timeframes
These studies measure different things. Read each figure with its geography, time period, and definition.
50–55% could be reshaped
BCG’s April 2026 model estimates that this share of U.S. jobs could change substantially over the following two to three years—approximately 2028–2029. Reshaping means the job remains but the work changes.
BCG analysis ↗10–15% potentially vulnerable
The same BCG analysis estimates potential U.S. job elimination over four to five years, roughly 2030–2031, possibly later. This is a model of potential impact, not an unemployment forecast or confirmed layoffs. Adoption and demand affect the outcome.
BCG methodology and limitations ↗92 million displaced by 2030
WEF’s global 2025–2030 outlook projects 92 million jobs displaced and 170 million created across multiple economic trends. These totals are not AI-only. Its separate AI and information-processing estimates are approximately 9 million displaced and 11 million created.
WEF original report ↗54,836 AI-cited cuts in 2025
Challenger, Gray & Christmas counted this many announced U.S. job cuts where employers cited AI during 2025. These are announced plans and employer-attributed reasons, not independently established AI causation or a count of completed layoffs.
Challenger year-end report ↗Professional and physical work
What about doctors, lawyers, drivers, and factories?
Doctors and medical specialists
Yes—physician specialties are included. Radiologists have high relative AI exposure, with 3.4% employment growth projected for 2025–2035. Family medicine physicians are also high exposure, with 3.3% growth. Surgeons and other specialists have their own entries.
Exposure identifies work AI may assist or perform; it does not establish when doctors will be replaced.
BLS occupation data ↗Lawyers and legal support
Lawyers have very high relative AI exposure, while BLS projects 4.7% employment growth for 2025–2035. Legal assistants, legal secretaries, judges, and other legal roles appear separately.
A high exposure category can coexist with growing demand. It is not a percentage of lawyers expected to lose their jobs.
BLS occupation data ↗Delivery and truck drivers
Light truck drivers have moderate AI exposure and 6.3% projected growth; heavy truck drivers have moderate exposure and 3.8% growth through 2035.
Physical automation is also real: Aurora announced commercial driverless trucking in Texas in May 2025, and Wing announced a Walmart drone-delivery expansion in 2026. These are deployments and company plans, not dates when all driving jobs end.
Factory and warehouse workers
Production jobs differ sharply. BLS projects 17.0% fewer engine and other machine assemblers, 9.2% fewer tool and die makers, and 1.0% more machinists during 2025–2035. These are total employment forecasts, not AI-only reductions.
Robotics can affect physical tasks even where the language-focused AI exposure score is low. Amazon’s 2025 Vulcan announcement describes robotic picking and stowing; it supplies no nationwide job-elimination date.
How to read the outlook
The list is ordered by relative AI exposure: Very high, High, Moderate, then Low. Within each category, the largest projected employment declines appear first, followed by growth. These categories are not individual job-loss probabilities.
When are these changes projected to happen?
The BLS employment forecast compares 2025 with 2035. It does not identify the year within that decade when a reduction will happen. The separate World Economic Forum outlook covers global employer expectations for 2025–2030. Neither supplies a reliable date when each occupation will end. Read the WEF outlook ↗
What do Low, Moderate, High, and Very high mean?
They are BLS’s relative AI exposure categories, combining five research datasets. They describe the potential or observed use of AI in occupational tasks, not replacement probabilities, wage effects, or guaranteed productivity gains. The inputs mainly concern language-based AI; physical robotics is not comprehensively measured. Low exposure does not mean immune to automation. Read BLS’s methodology ↗
Is this really the full list?
It includes all 831 detailed occupations in the BLS AI exposure dataset, across all 22 major job groups—not just occupations expected to decline. BLS groups many everyday job titles into broader occupations, so it is not a list of every employer-specific title. Search synonyms or browse the relevant group when your exact title is absent.
Can a job grow even when AI performs some of its tasks?
Yes. Demand, population needs, costs, and productivity can change together. Employers may redesign work instead of eliminating an occupation. The ILO’s 2025 research emphasizes task transformation as more likely than complete replacement. A shrinking occupation can also continue hiring to replace workers who retire or leave. Read the ILO study ↗
Plan your next step with evidence
Identify which parts of your work are routine, which depend on human judgment or physical presence, and what customers will still pay for. Compare local demand and training needs before changing careers or starting a business.
Explore 100 business ideas →Sources & scope
The complete list uses the BLS AI exposure and 2025–2035 employment projections file, with categories explained in the BLS methodology. The projection release is dated August 27, 2026. Data refer to the United States; WEF and ILO provide separate global context.
The exposure index combines theoretical and observed measures with different historical reference periods; it is not a real-time audit of every AI tool. Company announcements illustrate physical automation and are not independent forecasts of occupation-wide job losses.