Artificial intelligence and autonomous digital systems could force roughly 11 million American workers—equivalent to approximately 7% of the total United States workforce—to leave their current professions and transition into entirely new career fields by 2035, according to a comprehensive labor study published by the McKinsey Global Institute (MGI).
The report, titled “Workforce in Motion: Skills and Pathways to Future Jobs in the United States,” outlines a structural paradox defining the coming decade: while the US economy is projected to create more net jobs than automation destroys, the velocity and friction of the transition will create an unprecedented labor-mobility bottleneck. With only one in seven displaced workers possessing a direct, unhindered bridge to growing occupations, the study warns that the core economic challenge of the AI era is mobility, not scarcity.
Key takeaways
- 11 million occupational shifts: Between 6 million and 16 million US workers (with a baseline estimate of 11 million) in declining occupational categories will need to transition into completely different career tracks by 2035.
- Mobility over scarcity: Automation is projected to displace tasks equivalent to 36 million jobs, but economic expansion, infrastructure upgrades, and emerging AI sectors will create roughly 40 million new positions—generating a net positive jobs balance (+4 million) overshadowed by an acute reallocation crisis.
- Triple the historical transition speed: An estimated 770,000 workers will need to switch occupational fields each year through 2035—more than triple the historical annual rate of career migration in the US economy.
- The credential wall: While every displaced worker has potential pathways into growing industries, only 14% (1 in 7) can make the leap without substantial retraining, as roughly 85% of expanding jobs require formal certifications, licenses, or post-secondary credentials.
- Pervasive task reinvention: More than 70% of all US workers will see their day-to-day responsibilities fundamentally reinvented, with automated software taking on at least 15% of their daily working hours even if they remain in the same job title.
- Wage polarization risks: Over 70% of declining roles are concentrated in the bottom two income quintiles (such as administrative support, basic bookkeeping, and retail sales), while 60% of growing positions sit in the top two income brackets (such as healthcare specialists, data engineers, and construction supervisors).
Why the AI transition is a mobility crisis, not a scarcity crisis
Public debate surrounding artificial intelligence often focuses on the fear of mass structural unemployment—the prospect that synthetic cognition will permanently reduce the aggregate volume of human labor required by modern industry.
However, McKinsey’s empirical modeling reveals that aggregate labor demand will remain resilient. Driven by an aging population requiring healthcare, multi-trillion-dollar investments in domestic manufacturing and green energy grids, and the expansion of the digital economy, the US will generate tens of millions of roles.
THE US LABOR BALANCE SHEET: 2024–2035 (MCKINSEY PROJECTIONS)
Work-Hours Displaced by Automation:
[████████████████████████████████████] ~36 Million Job Equivalents
New Labor Demand Generated by Economic Growth:
[████████████████████████████████████████] ~40 Million New Roles Created
│
▼
[ Net Jobs Balance: +4 Million Positions ] (The Economy Is Not Running Out of Work)
│
▼
[ The Underlying Friction: 11 Million Workers Must Change Careers Completely ]
The friction lies in the mismatch between who is displaced and where the new jobs are being created:
- The Displaced: Concentrated heavily in routine, predictable digital and physical workflows—customer service reps, billing clerks, paralegal researchers, and entry-level coders.
- The Demanded: Concentrated in direct interpersonal care, specialized trade infrastructure, advanced systems management, and physical-world execution.
A displaced billing clerk cannot seamlessly transition into an operating-room surgical technician or an electrical grid engineer without years of structured retraining, credentialing, and income support. This mismatch forms the core of McKinsey’s warning: workers are not being shut out of work; they are trapped behind skills and licensing barriers.
Sector exposure: Where work hours will be automated
The report establishes that automation will not impact sectors evenly. By evaluating thousands of detailed workplace tasks across hundreds of occupations documented by the US Bureau of Labor Statistics (BLS), McKinsey identified the expected adoption of automation across working hours by 2035:
| Major Occupational Group | Expected Automation of Current Hours (2035) | Primary Driving Technologies | Core Operational Impact |
| Office & Administrative Support | 80% | LLMs, Agentic AI, Automated CRM | Drafting memos, scheduling, billing data entry eliminated |
| Technology & Data Analytics | 70% | Automated Code Generators, Neural IDEs | Standard software testing, boilerplate scripting automated |
| Retail, Sales & Customer Care | 68% | Autonomous Conversational Agents | Automated tier-1 support, dynamic digital storefronts |
| Transportation & Warehousing | 48% | Computer Vision, Robotics, Yard Logistics | Sorting hubs automated; long-haul autonomous piloting |
| Production & Manufacturing | 45% | Industrial Cobots, Vision Inspection | Assembly lines automated; maintenance shifts to robotics |
| Healthcare Professionals | 26% | Clinical Diagnostics Co-pilots | Administrative chart work cut; physical bedside care intact |
| Public Safety, Police & Fire | 28% | Video Analytics, Predictive Dispatch | Direct physical emergency response remains manual |
Source: Compiled from McKinsey Global Institute (MGI) research and BLS occupational task distributions.
Administrative support faces the highest degree of disruption, with four out of five working hours (80%) susceptible to automation by 2035. Conversely, roles requiring hands-on dexterity, physical presence, unpredictable environmental navigation, and deep interpersonal empathy—most notably bedside healthcare—exhibit the highest insulation from direct displacement.
PERCENTAGE OF WORK HOURS AUTOMATABLE BY 2035:
Office & Administrative Support: [████████████████████████████████] 80%
Technology & Analytics: [████████████████████████] 70%
Retail & Sales: [██████████████████████] 68%
Transportation & Logistics: [███████████████] 48%
Production & Manufacturing: [██████████████] 45%
Public Safety & Emergency: [█████████] 28%
Healthcare Professionals: [████████] 26%
The three-tier pathway trap: Credentials as structural barricades
One of the study’s central insights is the classification of occupational transition pathways into three distinct tiers:
TRANSITION PATHWAYS FOR THE 11 MILLION WORKERS:
[ Direct & Paved Pathways ] ─────► Only ~14% (1 in 7 Workers)
Can transfer immediately with minimal or zero retraining.
[ Twisted Pathways ] ────────────► ~38% of Displaced Workforce
Have foundational skills but face arbitrary licensing/credential barriers.
[ Unpaved & Broken Pathways ] ───► ~48% (Nearly Half of Displaced Workers)
Require multi-year education, complete reskilling, and debt financing.
- Direct Pathways (14%): Only one in seven workers currently enjoys an open bridge from a declining occupation to an expanding one where their existing experience translates directly without mandatory schooling or income loss.
- Twisted Pathways (38%): Workers have the practical cognitive abilities to perform the new role, but state occupational licensing boards, degree requirements, or credential monopolies prevent them from being hired without undergoing costly certification.
- Unpaved Pathways (48%): Nearly half of all transitioning workers face an educational cliff. To enter growing occupations in healthcare, data systems, or specialized engineering, they must undergo two to four years of full-time training, an impossibility for low-wage earners supporting families without subsidized living stipends.
Because 85% of growing jobs require credentials, the paper-qualification ceiling threatens to lock millions of displaced service and clerical workers into persistent underemployment.
Task reinvention: The 70% of workers who won’t lose their jobs
For the majority of the US workforce, artificial intelligence will not cause pink slips or career changes; instead, it will alter the nature of their daily job descriptions.
McKinsey projects that more than 70% of all US workers will experience substantial task reinvention, defined as having more than 15% of their working hours transferred to automated AI workflows.
THE REINVENTED KNOWLEDGE WORKER (TASK DISTRIBUTION):
PRE-AI WORKDAY (8 HOURS):
[ Data Gathering & Processing: 4 Hrs ] [ Report Writing: 2.5 Hrs ] [ Strategy & Client Care: 1.5 Hrs ]
AI-REINVENTED WORKDAY (8 HOURS):
[ AI Generation & Audit: 1.5 Hrs ] [ Strategy, Synthesis & High-Touch Human Advisory: 6.5 Hrs ]
Rather than executing routine analyses, writing first-draft summaries, or manually managing customer ticket queues, workers will transition into supervisory and editorial roles—reviewing machine outputs, validating synthetic conclusions against compliance guidelines, and handling edge cases requiring human emotional intelligence.
The emerging skill taxonomy: Essential, Enabling, and Empowering
To help employers and workers navigate this transition, McKinsey categorizes human workforce competencies into three functional categories:
- Essential Skills: Foundational baseline competencies that widen career alternatives across multiple fields—including digital literacy, structured communication, basic statistical reasoning, and collaborative problem-solving.
- Enabling Skills: High-leverage operational skills that unlock entry into higher wage tiers—such as specialized project management, programming architectures, regulatory compliance, and clinical patient care.
- Empowering Skills: Psychological and cognitive traits that allow workers to endure continuous technological disruption.
The report documents an unprecedented surge in market demand for empowering skills across corporate job postings since the commercial rollout of generative AI in 2022:
GROWTH IN DEMAND FOR EMPOWERING SKILLS (SINCE 2022):
Demand for Resilience, Curiosity & Learning Agility: [███] 3x Increase
Demand for Workplace Adaptability: [█████] 5x Increase
Demand for Practical AI Fluency: [███████████] 11x Increase
The data shows that employers are increasingly valuing cognitive adaptability—the demonstrated capacity to continuously learn new digital tooling—over static historical domain expertise.
Macroeconomic and policy ramifications: Rebuilding the labor transition infrastructure
The prospect of 11 million workers needing to cross career boundaries by 2035 exposes structural weaknesses in modern public-policy frameworks:
1. Reforming occupational licensing and credentialism
State governments across the US maintain extensive licensing requirements for mid-tier professions, ranging from medical records management to commercial logistics. If states do not reform these rules by recognizing competency-based assessments and practical apprenticeships in place of formal university degrees, labor bottlenecks will worsen while displaced workers remain sidelined.
2. Lifelong reskilling accounts and wage insurance
Traditional unemployment insurance models were engineered for temporary cyclical recessions—paying workers a short-term stipend while they wait for their old factory or office to rehire them. In an AI-driven economy, the old jobs do not return. Economists argue that governments and enterprise consortiums must establish transferable lifelong learning accounts, subsidized apprenticeships, and transitional wage-insurance pools to support workers as they retrain for high-growth sectors.
3. Deepening geographic disparities
Expanding roles in technology, modern manufacturing, and specialized clinical care are concentrating in high-productivity metropolitan clusters and federally designated semiconductor and green-energy hubs (such as Texas, Arizona, North Carolina, and Ohio). Conversely, clerical and administrative displacement will be felt across every town and municipality nationwide, threatening to widen rural-urban wealth divides if remote-work bridges and localized vocational training hubs are not established.
Global implications: What the McKinsey findings mean for India
While the McKinsey study focuses on the United States, its conclusions carry direct implications for India’s digital economy and corporate workforce:
- Back-Office and BPO Transformation: India’s vast business process management (BPM) and IT-enabled services (ITeS) sectors employ millions of professionals performing administrative support, customer helpdesk ticketing, and data reconciliation for US and European corporations. With 80% of office and administrative work hours automatable by 2035, Indian services giants (such as TCS, Infosys, Wipro, and Genpact) face an imperative to transition contract models from billing manual man-hours toward deploying and maintaining enterprise AI agents.
- The “Human Economy” Export Opportunity: As the US and other developed Western nations face shortages in healthcare workers, geriatric care, and physical infrastructure engineers due to aging populations, India’s demographic dividend positions it as an essential global talent supplier—provided domestic educational institutions upgrade vocational training to match international clinical and technical accreditation standards.
- Engineering Workflow Reinvention: With 70% of tech and analytics hours automatable, Indian software developers are transitioning from writing basic boilerplate code to orchestrating complex AI systems, demanding accelerated domestic upskilling in system architecture and AI safety governance.
What could happen next
- Enterprise workforce audits: Fortune 500 corporations will accelerate formal workforce skills audits heading into 2027, mapping internal job descriptions to identify roles facing high automation exposure.
- Federal workforce legislation: US lawmakers on both sides of the aisle are expected to evaluate legislative proposals expanding Pell Grants to short-term vocational credentials and offering tax credits to companies that invest in employee reskilling.
- Corporate-university apprenticeships: Major technology firms and healthcare networks will expand direct-hire apprenticeship models, bypassing traditional four-year degree programs to recruit and train transitioning clerical workers directly.
Frequently asked questions
Will AI cause net job losses in the United States by 2035?
No, according to McKinsey. The US economy is projected to create approximately 40 million new jobs by 2035, exceeding the 36 million job equivalents displaced by automation. However, roughly 11 million workers in declining occupations will need to switch into entirely different career fields, creating a significant transition challenge.
Which jobs are most at risk of displacement from AI?
Office and administrative support roles face the highest exposure, with up to 80% of current work hours automatable by 2035. Other highly exposed categories include technology and data analytics (70%), retail sales and customer support (68%), and transportation logistics (48%).
Which sectors will see the highest job growth?
The highest job growth will occur in healthcare (driven by an aging population), construction and infrastructure modernization, renewable energy installations, and management roles requiring high-touch interpersonal decision-making.
Why is it difficult for displaced workers to move into growing jobs?
Roughly 85% of growing professions require formal certifications, licenses, or post-secondary credentials. Currently, only one in seven displaced workers (14%) has a direct pathway to transition without substantial time- and capital-intensive retraining.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
