Microsoft study names 20 jobs least likely to be hit by AI — healthcare and blue‑collar roles lead the list
A new Microsoft Research report maps where generative AI fits into everyday jobs, finding hands‑on healthcare and blue‑collar roles score lowest in AI applicability — offering career planning insight on occupations most resilient to current LLMs.
- Microsoft computed an “AI applicability score” from 200,000 anonymized Bing Copilot conversations to measure how well LLMs map to common job tasks. Source: Source Source
- Healthcare support and blue‑collar trades score lowest — these roles require physical presence, situational judgment, and manual dexterity that LLMs can’t replicate. Source: Source Source
- High applicability ≠ guaranteed job loss — the study measures LLM usefulness for tasks, not total automation risk from other AI systems or robotics. Source: Source
What Microsoft studied and how
Approach: Microsoft matched anonymized Bing Copilot conversations to occupational task data and computed an AI applicability score that rates how well current LLMs map to everyday job tasks. The method focuses on language and information work, not robotics or sensor-driven systems. Source: Source Source
“This test does not predict job loss directly; it shows where current generative AI is best suited to support workers.”
Tasks like writing, summarizing, or translating score high because LLMs already perform those tasks effectively; hands‑on and site‑specific tasks score low because they require physical presence, real‑time judgment, and unpredictable environments. Source: Source
Top 10 least affected occupations (lowest AI applicability)
Microsoft’s list of occupations least likely to be affected by generative AI highlights manual, site‑specific, and operational jobs. These are presented exactly as identified in the report:
- Dredge operators (operate equipment to deepen waterways)
- Bridge and lock tenders (manage bridge and lock operations)
- Water treatment plant and system operators (monitor and control water purification)
- Foundry mold and coremakers (fabricate molds for metal casting)
- Rail‑track laying and maintenance equipment operators (maintain railway infrastructure)
- Pile driver operators (drive piles for construction foundations)
- Floor sanders and finishers (refinish wood floors manually)
- Orderlies (assist in patient handling and transport in healthcare)
- Motorboat operators (navigate and operate boats)
- Logging equipment operators (handle machinery for timber harvesting)
Broader patterns: why healthcare and blue‑collar jobs score low
Healthcare support roles (for example, home health aides, personal care aides, and orderlies) typically require direct physical care, touch, and real‑time human judgment. Those elements are difficult for LLMs to assist with in a hands‑on way, while many healthcare support jobs remain among the fastest‑growing occupations. Source: Source
Blue‑collar trades in construction, manufacturing, utilities, and transportation involve heavy equipment, outdoor variability, safety decisions, manual dexterity, and rapid responses — constraints that reduce LLM usefulness. The report notes other AI forms (robots, sensors) could affect some roles over time. Source: Source
Where AI looks most applicable
Work centered on communication, writing, research, and repetitive information tasks shows higher AI applicability. Occupations such as translators, writers, customer service and sales representatives, and many office jobs align with LLM strengths and may use AI as a productivity tool rather than a straight replacement. Source: Source
Limitations and what the scores mean
Key caveats in the report:
- LLM applicability only: The measure does not capture impact from non‑LLM AI such as industrial controls, sensors, or process automation. Source: Source
- Applicability ≠ displacement: A high score means AI can assist with tasks, not that jobs will be fully replaced; roles may evolve, be augmented, or spawn new positions tied to AI. Source: Source
- Data reflects current tool use: The dataset relies on Bing Copilot conversations and may change as usage patterns and tools evolve. Source: Source
Microsoft also highlights emerging AI‑related roles — AI trainers, AI governance roles, and Chief AI Officers — that may offer alternative career paths while keeping hands‑on trades steady. Source: Source Source
Implications for Alamogordo, New Mexico
Economic impact: Alamogordo’s economy includes healthcare services, construction firms, utilities, and light manufacturing tied to nearby military bases. The Microsoft findings suggest jobs least impacted by AI — such as orderlies, equipment operators, and maintenance trades — are likely to remain in demand. Encouraging local training for blue‑collar and healthcare support roles can help sustain wages and hiring. Source: Source
Political and workforce planning: For readers focused on economic stability and local control, the study supports investment in trade education, community college programs, apprenticeships, and employer partnerships to build a resilient workforce. Source: Source
Social and community effects: Many low‑applicability jobs involve direct care and service. Preserving and improving pathways into these fields supports care for veterans, seniors, and neighbors while protecting family‑supporting local jobs. Source: Source
Practical advice for residents:
- Consider training or upskilling in trades, heavy equipment operation, or healthcare support roles.
- Support local apprenticeship programs and hire‑local initiatives to strengthen workforce pipelines.
- Office workers should learn to use AI tools as productivity aids and seek roles emphasizing human judgment, inspection, or supervision. Source: Source Source
Where to read the research
- Microsoft Research paper and notes
- Microsoft blog discussion and caveats
- ArXiv preprint
- Reporting on the study (Fortune)
- Microsoft work trends and AI employment context
- Microsoft Work Trend Index 2025
Summary: This analysis highlights practical pathways for Alamogordo workers and leaders: reinforce training in blue‑collar trades and healthcare support, and prepare office workers to augment their roles with AI rather than expect direct replacement.
