McKinsey Sees 36 Million U.S. Jobs’ Worth of AI-Driven Labour Shift
A new McKinsey analysis projects major task and occupation changes by 2035, while other evidence finds no broad AI jobs shock yet.
McKinsey Global Institute estimates that automation and AI could reduce U.S. labour demand equivalent to 36 million jobs by 2035, a 21% reduction. The report also projects demand for more than 40 million new jobs, meaning the change would involve substantial occupational movement rather than only job losses.
About 11 million workers—roughly 7% of the workforce—may need to switch occupations, while 25 million could remain in roles with changed tasks. Office and administrative support is among the most exposed areas, with 81% of work hours affected and demand reduced by 38%. More than 75% of workers needing occupational shifts are in administrative support, retail and sales, and transport and logistics, according to the report as reported by CNBC.
The findings point to a long-term restructuring risk, not evidence of immediate economy-wide displacement. The Yale Budget Lab and displacement.ai said on 15 October that broad AI-driven labour-market disruption was not yet clear, while CEPR found that AI often changes hiring and task allocation before producing mass redundancies.
For employers: treat exposure forecasts as a reason to map tasks and redesign work—not as automatic grounds for dismissal. If roles may disappear, document the operational change, consult affected employees before deciding, assess genuine redeployment options, and only then determine whether a lawful genuine redundancy is required.
- McKinsey projects a U.S. labour-demand reduction equivalent to 36 million jobs by 2035.
- More than 40 million new jobs could be created, alongside occupational shifts for about 11 million workers.
- Administrative, retail, sales, and transport roles account for most projected occupational switching.
- Other 2026 evidence finds no clear broad AI-driven labour-market disruption yet.
For employers: Use AI exposure forecasts to map changed tasks and redeployment options, not to predetermine redundancies. Document the genuine operational change, consult before deciding, assess suitable redeployment, and only then determine whether a lawful genuine redundancy is necessary.
Sources: McKinsey report: Automation will affect 36 million jobs by 2035 · Yale Budget Lab / displacement.ai labour-market updates · CEPR: Hiding in plain sight—AI is adding jobs and removing them at the same time
General information for employers, not legal advice. Reporting summarised from the linked sources; confirm details with the primary source.