36 Million Jobs Could Be Replaced by AI: The 10 Most Affected Fields
Jakarta, CNBC Indonesia — Advances in artificial intelligence (AI) and automation technology are expected to transform the world of work on a large scale over the coming decade. Demand for workers in some occupations could decline as more tasks become possible for technology to perform.
A McKinsey Global Institute report published on 29 September 2026 estimates that around 36 million jobs — equivalent to 21% of US labour demand — could be affected by reduced staffing needs due to automation by 2035.
This does not mean all those jobs will disappear. An estimated 25 million workers could remain in their occupations, although their duties and working methods are likely to change.
Meanwhile, around 11 million workers, or 7% of the US workforce, may need to move into other occupations over the next decade. At the same time, AI is projected to create more than 40 million new jobs in the US over the same period.
“The good news is that labour demand will be higher than the decline,” McKinsey Global Institute director Tanguy Catlin said, according to CNBC Make It on Monday (12 October 2026).
Catlin cautioned, however, that the transition would not be easy for all workers, particularly those who have to move into entirely different fields.
According to the McKinsey Global Institute report, the following are the 10 occupational groups expected to be most affected by the adoption of automation by 2035.
- Office and administrative work
Number of workers: 18 million
Share of working hours potentially affected by automation: 81%
Decline in labour demand: 38%
- Agriculture, fishing and forestry
Number of workers: 400,000
Share of working hours potentially affected by automation: 76%
Decline in labour demand: 43%
- Technology and data analysis
Number of workers: 5.2 million
Share of working hours potentially affected by automation: 70%
Decline in labour demand: 22%
- Manufacturing and production
Number of workers: 8.5 million
Share of working hours potentially affected by automation: 68%
Decline in labour demand: 25%
- Retail and sales
Number of workers: 13.3 million
Share of working hours potentially affected by automation: 67%
Decline in labour demand: 27%
- Business and finance
Number of workers: 10.5 million
Share of working hours potentially affected by automation: 62%
Decline in labour demand: 18%
- Food and drink services
Number of workers: 13.7 million
Share of working hours potentially affected by automation: 60%
Decline in labour demand: 31%
- Arts, media and entertainment
Number of workers: 1.9 million
Share of working hours potentially affected by automation: 58%
Decline in labour demand: 13%
- Transport and logistics
Number of workers: 13.8 million
Share of working hours potentially affected by automation: 55%
Decline in labour demand: 30%
- Installation and repair
Number of workers: 6.1 million
Share of working hours potentially affected by automation: 54%
Decline in labour demand: 15%
AI does not always replace workers
Catlin explained that a high potential for automation in a particular occupation does not necessarily mean companies will cut staff numbers. Some professions, he said, are instead expected to undergo major changes in how work is done, as AI takes over certain tasks while people continue to handle work requiring judgement and decision-making.
For example, around 47% of working hours in the legal sector could be affected by automation. However, the projected decline in labour demand in the sector is only 18%.
Catlin said lawyers previously spent a great deal of time conducting legal research. Some of this work can now be assisted by AI, giving lawyers more time to handle tasks that require human judgement.
“Your job title or position does not change, but the way you spend your time at work changes dramatically,” he said.
He also stressed the difference between work that can theoretically be automated and work that will actually be replaced by technology. The McKinsey report says technology determines what is possible, while company management decisions determine how quickly those changes are implemented.
As a result, the number of workers who may have to change occupations is estimated to range from 6 million to 16 million, with a baseline estimate of 11 million.
Millions of workers may struggle to find new jobs
The report also finds that job transitions resulting from automation could pose a major challenge for millions of workers. Of the estimated 11 million workers who may need to change occupations, only one in seven has a relatively straightforward path into a new job.
That means they could move into fields requiring similar skills, with easily obtained certification requirements and salaries equal to or higher than those in their previous jobs. By contrast, nearly 45% of workers are expected to face a much more difficult transition.
They may need to relocate, learn significantly different skills or obtain additional certifications. The new jobs available may not offer salaries as high as their previous roles.
“To obtain those certifications, workers may have to invest time and money, and this will be one of the obstacles,” Catlin said.
For 11 million workers to move into new occupations over the next decade, an estimated 770,000 workers would need to change occupational groups each year. That is about 3.6 times the historical average rate of occupational transition.
Skills to develop to remain resilient to AI
Amid rapid technological advances, Catlin advised workers to strengthen skills that are not easily replaced by automation. These include problem-solving, interpersonal communication, teamwork, critical thinking, curiosity and a willingness to keep learning.
These skills could help workers secure better job opportunities and improve their ability to adapt to changes in the world of work. Catlin warned that technological development is likely to continue at a rapid pace, so workers need to make continuous skills development a habit throughout their careers.
“The pace of change will never be slower than it is today. We therefore need a workforce that can adapt, keep developing its skills and continue learning,” he said.
Source: jawawa.id (https://jawawa.id/newsitem/36-million-jobs-could-be-replaced-by-ai-the-10-most-affected-fields-1791777100)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/lifestyle/20261012092739-33-775237/36-juta-pekerjaan-digantikan-ai-ini-10-bidang-paling-terdampak), translated from Indonesian