Microsoft study names 20 jobs least likely to be shaken by AI — blue‑collar and healthcare roles top the list
Microsoft research identifies U.S. occupations least likely to be impacted by generative AI, finding manual, in‑person work—dredge operators, orderlies, plumbers—and many healthcare and trade roles show the lowest AI applicability scores, while knowledge work is most exposed.
- Microsoft measured an “AI applicability score” by mapping ~200,000 anonymized Bing Copilot chats to O*NET tasks to see where LLMs can assist. See the Microsoft research paper.
- Blue‑collar and healthcare jobs rank lowest because they require physical presence, touch, mobility and unpredictable judgment that limit LLM help.
- Applicability ≠ displacement: the study shows where AI can assist, not where jobs will be lost; read Microsoft’s clarification applicability vs. displacement.
- Local relevance: places like Walla Walla with farms, wineries, construction and health care may see steadier demand for hands‑on roles.
The 20 jobs least likely to be affected by generative AI
Microsoft publicly highlighted a top 10 of occupations with the lowest AI applicability; the next 10 below follow the same low‑applicability pattern described in the paper and reporting but were inferred rather than formally released as a numbered top‑20 by Microsoft.
- Dredge operators — heavy equipment and on‑site work. Source
- Bridge and lock tenders — physical, location‑bound tasks. Source
- Water treatment plant and system operators — monitoring and equipment work. Source
- Foundry mold and coremakers — hands‑on manufacturing. Source
- Rail‑track laying and maintenance equipment operators — on‑site heavy machinery. Source
- Pile driver operators — construction equipment specialists. Source
- Floor sanders and finishers — skilled manual trade work. Source
- Orderlies — hands‑on patient care and movement. Source
- Motorboat operators — navigation and on‑water work. Source
- Logging equipment operators — heavy machinery in remote sites. Source
- Home health and personal care aides — close personal contact and hands‑on tasks. Source
- Nursing assistants — patient care duties that require presence and touch. Source
- Food preparation and serving workers (non‑automated lines) — manual food work and on‑site service. Source
- Construction laborers — varied manual tasks and machine operation on site. Source
- Roofers — outdoor, physical trade work. Source
- Plumbers, pipefitters, and steamfitters — skilled troubleshooting in physical environments. Source / arXiv
- Electricians — on‑site diagnostics and installations. Source
- Carpenters — hands‑on fabrication and installation. Source
- Licensed practical/vocational nurses (LPN/LVNs) — routine clinical work and direct care. Source
- Dental assistants — chairside, manual and patient‑facing tasks. Source
How Microsoft measured “AI risk”
Microsoft’s team labeled tasks drawn from roughly 200,000 privacy‑scrubbed Bing Copilot chats (writing, researching, summarizing, coding and more) and matched those activities to O*NET task descriptions to build an “AI applicability score”. The method shows how often a Copilot‑style LLM could assist. Read the full methodology in the Microsoft research paper and the arXiv preprint.
“The study measures where LLMs can add value. It does not predict layoffs or the timing of change.” — Microsoft clarification on applicability vs. displacement
Why healthcare and blue‑collar jobs look safer
The pattern is straightforward: jobs that need a person on site, performing physical or direct‑care work, score low for LLM help. Healthcare support roles and many trades require touch, mobility, judgment in unpredictable settings, and close human interaction — traits Microsoft and reporting cite as reasons these occupations have low applicability. See the Fortune summary: Fortune summary.
Caveats and what the study does not say
- LLM focus: the research centers on generative language models; other technologies — robotics, computer vision and automation — could affect manual jobs later. See arXiv.
- Assistance, not layoffs: every job has some tasks AI can help with; the score measures assistance potential, not job losses. Read Microsoft’s note on applicability vs. displacement.
- Additional items 11–20: the second set of ten occupations listed above were inferred from Microsoft’s described pattern and reporting, not formally published as a top‑20 by Microsoft; treat them as consistent with the study’s findings.
Implications for Walla Walla, Washington
Economic impact
Walla Walla’s economy depends on hands‑on work: farms, food processing, wineries, construction, and health care services. The Microsoft AI study suggests jobs rooted in physical work and direct care are less likely to see LLM assistance today, meaning local blue‑collar and healthcare jobs — from plumbers and electricians to nursing assistants and home health aides — should remain in steady demand in the near term. See the Microsoft research paper and the Fortune summary.
Political consequences
Local leaders may use the study to argue for expanded vocational training and apprenticeships. For conservative audiences this can mean emphasizing private‑sector partnerships and market‑driven training at community colleges. Walla Walla Community College and local trade programs can highlight these career pathways. See methodology notes: arXiv.
Social effects
Health care support roles figure prominently in the low‑applicability group. As the region ages and demand for patient care grows, home health aides, orderlies, LPNs and nursing assistants will remain important — supporting stable jobs for residents who choose direct care careers over office roles.
Cultural relevance
Skilled trades and hands‑on work are central to Walla Walla’s identity. Farmers, vintners, craftspeople and construction workers form the backbone of the local economy; the Microsoft findings validate the social and cultural value of trades that require real‑world presence and craftsmanship.
Practical applications for residents
- Job seekers: consider training in trades, health care support, or equipment operation — paths with low AI applicability today. See the Microsoft research paper.
- Employers: invest in on‑the‑job training and apprenticeships to fill hands‑on roles; emphasize these jobs’ enduring necessity.
- Policymakers and educators: strengthen community college and technical programs and promote partnerships with private employers to create clear career ladders; local institutions like Walla Walla Community College can be focal points.
