🔍 Read the full analysis: AI At Work: A Look At Three Kinds Of Unemployment on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A Nov. 16, 2026, article from ThorstenMeyerAI.com explains frictional, structural and cyclical unemployment and places technological unemployment within the structural category. It reports 4.1% US unemployment in October, alongside a rising share of people searching for work long term, but offers no evidence that AI has caused a broad rise in joblessness.
ThorstenMeyerAI.com published an explainer on Nov. 16 on how economists classify unemployment and whether artificial intelligence may be producing the technological unemployment John Maynard Keynes described in 1930. The article reports that the US unemployment rate was 4.1% in October 2026, while more than one in four unemployed people had been searching for at least six months; those figures describe labor-market conditions but do not establish that AI caused them.
The article identifies three standard categories: frictional unemployment, when people are between jobs or entering the job market; structural unemployment, when workers’ skills or locations do not match available jobs; and cyclical unemployment, when economic downturns reduce demand for workers. It says economists generally classify technological unemployment—the loss of work when machines or software take over tasks—as a form of structural unemployment.
Using the US Bureau of Labor Statistics’ definition, the article says a person is counted as unemployed if they had no job during the reference week, could have started work and actively looked for a job in the preceding four weeks. People who want work but have stopped searching are not counted in the headline unemployment rate. The article reports an October median unemployment spell of 11.9 weeks and says 27.6% of unemployed Americans had searched for at least 27 weeks.
For comparison, the article cites Germany’s Federal Employment Agency, which reported that 36% of registered unemployed people—1,063,000 people—had been out of work for more than a year. It cautions that the German and US measures use different definitions, so their levels should not be directly compared. The article also recalls that the US unemployment survey began in 1940, after the Great Depression increased demand for regular statistics.
Reading AI Through Jobless Data
The distinction matters because the headline unemployment rate is not a complete account of labor-market strain. It excludes people who have stopped looking, while long job searches can point to difficulties finding a match between workers and vacancies. The October figures cited in the article therefore offer different signals about the labor market: an overall rate of 4.1% and a substantial share of job seekers with lengthy spells of unemployment.
They do not, on their own, show whether AI is eliminating jobs, changing tasks within existing jobs or creating new work. A rise in long-term unemployment can have multiple causes, including economic conditions and mismatches between workers and open positions. For workers, employers and policymakers, identifying the cause matters: a recession-related decline may call for a different response than a persistent skills mismatch linked to changes in technology.
The article’s central point is that technological change is best understood through the structural-unemployment framework, while any claim that AI is causing mass unemployment needs evidence beyond a general jobless rate. The figures presented are a snapshot, not a causal analysis of AI’s effects.
Products Worth Considering

Bose Smart Ultra Dolby Atmos Soundbar with Voice Control, HDMI eARC, Black

Samsung Galaxy Tab A11+ Tablet (6GB RAM, 128GB Storage, Gray)

DJI Osmo 360 Motorcycle Handlebar POV Combo
As an affiliate, we earn on qualifying purchases.
From Keynes to the October Data
In a 1930 essay, Keynes described “technological unemployment” as a situation in which labor-saving inventions advance faster than workers can find new uses for their labor. He also called it a temporary phase of maladjustment. The article uses that historical framing to ask whether AI is now producing a similar mismatch, rather than treating the possibility as established fact.
The piece points to earlier debates over automation, including a 1966 US commission that concluded technology eliminates jobs, not work. It also cites research by David Autor and colleagues estimating that about 60% of US jobs in 2018 were in occupational titles that did not exist in 1940. That historical pattern suggests work changes over time, but it does not prove that workers displaced by a particular technology will quickly find comparable jobs.
The article says the regular US unemployment survey started in 1940 and that the BLS has published it since 1959. It cites historical estimates showing unemployment at 3.2% in 1929 and 24.9% in 1933. Those Depression-era figures were estimated retrospectively, because a monthly survey did not yet exist.
““Technology eliminates jobs, not work.””
— The 1966 US commission on technology and employment, as quoted by ThorstenMeyerAI.com
What the Figures Cannot Establish
The article does not provide data linking the October 2026 unemployment rate or long-term job searches to AI adoption. It does not identify which workers or industries may have lost jobs specifically because of AI, nor does it quantify new jobs created as AI changes work. Whether AI has increased unemployment overall remains unclear from the evidence presented.
The reported figures are also a snapshot, and the article does not give a longer time series for the October rate or the share of long-term unemployed workers. It says that the share searching for at least six months was more than one in four, up from about one in five in 2019, but the figures alone do not explain the increase. The US and German statistics are based on different definitions and should not be treated as directly comparable.
Watch Hiring and Job Spells
The article does not name a forthcoming study, policy decision or scheduled data release as its next milestone. To judge whether AI is changing unemployment, readers will need future labor-market data alongside evidence about hiring, layoffs, job vacancies, worker training and changes in specific occupations. Tracking those measures over time could help separate cyclical weakness from persistent skills mismatches and identify whether technology is a contributing cause.
For now, the October figures provide a current labor-market snapshot, while the article’s historical examples supply context rather than a forecast. Any stronger conclusion about AI’s employment effects will require evidence that connects changes in technology to outcomes for workers.
Key Questions
What are the three main types of unemployment?
Frictional unemployment occurs while people move between jobs or look for their first job. Structural unemployment reflects a mismatch between workers and available roles; cyclical unemployment rises when an economic downturn reduces demand for labor.
How does the article classify technological unemployment?
It says economists usually treat technological unemployment as structural unemployment, because technology can change which skills employers need or which tasks require workers.
Does the October 2026 US rate show that AI is causing job losses?
No. The article reports a 4.1% unemployment rate for October but does not link that figure to AI. The rate is a broad labor-market measure, not proof of a specific cause.
Who is counted as unemployed in the US measure described?
A person must have had no job during the reference week, been available to start work and actively looked for a job during the prior four weeks. People who want work but have stopped searching are not counted as unemployed in the headline rate.
What did the article report about long-term unemployment?
It says 27.6% of unemployed Americans had been looking for work for at least 27 weeks in October 2026. It also reports that Germany’s measure showed 36% of registered unemployed people had been without work for more than a year, while warning that the countries’ definitions differ.
Source: ThorstenMeyerAI.com
