Job Shifts Happen for Lots of Reasons, Not Just Technology Impact

The concern that artificial intelligence will eliminate jobs is rational: AI will eliminate some jobs; restructure others and often devalue existing jobs. That always happens with big technology and economic transitions. 

Still, at a micro level, workers often change jobs over time that are not caused in any significant way by technology. Many of us had "first jobs" in food service or retail at age 16, when it became legal for us to work, but then moved on to other roles in our later teens, then additional roles after education and training, with greater experience.

In other words, some amount of work life involves migration from some roles (food service, retail, hospitality) to something else as we age. Technology has little to do with such shifts. So we will have to be careful about assuming (incorrectly) too much about job shifts that happen because of applied AI.

For example, research from MIT’s David Autor and MIT PhD student Caroline Chin, Utrecht University’s Anna M. Salomons, and Northwestern’s Bryan Seegmiller finds that approximately 60 percent of jobs in the United States today didn’t exist in 1940, when more than 25 percent of work was in manufacturing and nearly 20 percent in farming and mining.

The rise of personal computing and the Internet destroyed 3.5 million jobs beginning in 1980, 

McKinsey consultants estimate, but then created 19.3 million jobs. There were fewer typists but more software developers, for example. 

More than 2.5 million app developer jobs were created, up from virtually none a few decades earlier. 

Spreadsheets might have displaced bookkeeping jobs, but increased demand for analytical functions. And sometimes the technology impact was paradoxical. Automated teller machines reduced the need for tellers at bank branches. But branches also made economic sense when it was possible to operate them with fewer tellers, so the number of bank branches grew. 

But not all job shifts are necessarily caused by automation. Lots of young people have jobs early in their work lives in food service, customer service or office support, but leave those jobs over time as their skills in the work force develop, as they finish education and training.

So we need to be careful about attributing "workers changing jobs" over time with the separate impact of automation or other technologies on job functions and numbers.

source: McKinsey 

AI is likely to accelerate existing job changes. Between 2019 and 2022, for example, according to the U.S. Bureau of Labor Statistics, about 8.6 million workers changed jobs in one field and moved to another. The losing industries included food service; customer service; office support and production work, all the sorts of jobs that AI should begin to replace or change. 

On the other hand, AI also is expected to affect higher-skilled, white-collar jobs more than previous technological shifts, potentially requiring up to 12 million occupational transitions by 2030, according to McKinsey consultants. In contrast to previous automation technologies, Generative AI excels in doing cognitive, non-routine tasks, a study by the Organization for Economic Cooperation and Development argues. 

Such job losses and gains are typical for big shifts in technology. 

Study Name	Date	Publisher	Key Conclusions
Technological Change and the Consequences of Job Loss	2023	American Economic Review	Technological change accounts for 45% of the decline in earnings after job loss. It requires workers to have new skills for newly created jobs in their prior occupation1.
Jobs Lost to Automation Statistics	2024	TeamStage	Up to 20 million manufacturing jobs could be lost to robots by 2030. 25% of jobs in the US are at high risk of automation. 1.6 manufacturing jobs are lost for every robot implemented2.
New Frontiers: The Origins and Content of New Work, 1940-2018	2024	Quarterly Journal of Economics	Since 1980, technology has replaced more U.S. jobs than it has generated. The study quantifies both job losses and gains due to technology3.
Future of Work and Automation	2021	Emory University AI	32-42% of jobs lost during the pandemic will not return. Roughly half of the tasks performed by today's workforce can be automated4.
World Economic Forum Study	2020	World Economic Forum	By the mid-2030s, 30% of jobs and 44% of workers with low levels of education will be at risk of automation4.
Goldman Sachs AI Impact Report	2024	Goldman Sachs	Generative AI could impact productivity growth by 1.5% annually over the next 10 years. About two-thirds of US jobs are exposed to some degree of automation by AI5.
Assessing the impact of technological change on similar occupations	2023	PMC	Job losses for nearly 60% of current employment will occur in low-skill, low-wage occupational groups. Many mid-skilled and highly skilled jobs are projected to grow in the next ten years4.
Goldman Sachs AI Impact Report	2023	Goldman Sachs	300 million jobs could be lost or diminished by AI. Generative AI could impact productivity growth by 1.5% annually over the next 10 years. About two-thirds of US jobs are exposed to some degree of automation by AI5.
OpenAI GPT Impact Study	2023	OpenAI	At least 80% of the U.S. labor force could have at least 10% of their work-related tasks affected by GPT. 19% of employees may see at least 50% of their work-related tasks impacted2.
McKinsey Global Institute AI Report	2023	McKinsey	Without generative AI, automation could take over tasks accounting for 21.5% of hours worked in the U.S. economy by 2030. With generative AI, that share increases to 29.5%2.