🔍 Read the full analysis: Fourth Sector Dynamics In AI Era on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Thorsten Meyer’s November 10, 2026, article examines the growth of knowledge and information work, sometimes called the quaternary or fourth sector, and its exposure to generative AI. Historical estimates show the sector’s growing share of U.S. employment and output, but exposure to AI does not establish that jobs will be lost; the article says broad U.S. employment data has not yet shown a clear AI effect.
Thorsten Meyer published a report on November 10 about the growth of knowledge and information work—the economy’s so-called fourth, or quaternary, sector—and its exposure to generative AI. Drawing on historical estimates by Fritz Machlup and Marc Porat, the article argues that office and professional work has become a central part of the economy, while stressing that AI exposure is not the same as job loss.
Meyer’s article, part of the series Work After Work, revisits Machlup’s 1962 estimate that the production and distribution of knowledge accounted for $136.4 billion, or “almost 29 per cent of adjusted GNP,” in the United States in 1958. That figure measured spending on knowledge-related activity, not the value of knowledge itself. The categories included education, research and development, communications media, information machines and information services.
Machlup’s estimate also showed how broad the category was: education accounted for 44% of the measured total, media 28%, information services 13%, research and development 8% and information machines 7%. The article attributes that breakdown to historian of statistics Benoît Godin’s tabulation. It reports that knowledge industries grew by 8.8% a year between 1947 and 1958, according to Machlup’s estimates.
Meyer then turns to Porat’s 1977 study for the U.S. Department of Commerce, which classified workers by their tasks rather than their employer’s industry. In Porat’s figures, information workers made up 5.8% of the U.S. workforce in 1860 and 46.4% in 1970, when they were the largest of four groups: agriculture, industry, services and information. Those are historical estimates, not current employment statistics.
Infographic 01 · The number to remember
Fritz Machlup's first count of the knowledge industries, by branch.
US spending on knowledge by branch, 1958 · share of $136.4 bn
Machlup measured what knowledge cost (spending, the way national accounts count), not what it was worth. “Almost 29 per cent of adjusted GNP” is his figure for 1958. The branch values are as tabulated by Godin (2008); the labour shares are Machlup’s estimate as summarised by his publisher.
Sources: F. Machlup, The Production and Distribution of Knowledge in the United States (Princeton University Press, 1962), as tabulated by B. Godin, “The Knowledge Economy: Fritz Machlup's Construction of a Synthetic Concept” (2008); Princeton University Press, book description.
Powered by Thorsten Meyer AI
Why AI Targets Information Work
The sector matters because many generative AI tools are designed to perform or assist with tasks associated with clerical, professional and managerial jobs: producing and processing text, handling information and supporting analysis. If these activities make up a large share of work, changes to them could affect employers and workers well beyond the technology industry.
Meyer cites estimates that higher-income and clerical occupations are among those most exposed to generative AI. Exposure describes potential contact with the technology, not a confirmed outcome: it does not establish whether a task will be automated, a worker’s role will change, or employment will fall. The distinction matters for readers interpreting forecasts about the future of work.
The article also reports that U.S. employment data has not yet shown a clear overall AI footprint, while younger workers in the most exposed jobs are falling behind peers. These observations point to a possible uneven effect, but the source does not supply the underlying measures, time period or comparison details for the young-worker finding. It should not be read as proof that AI caused the difference.
Popular right now
LG 77-Inch OLED evo C6 4K Smart TV
DREAME A3 AWD 1000 LiDAR Robot Lawn Mower
KitchenAid Artisan Mini Tilt-Head Stand Mixer, 3.5 Qt, Contour Silver
As an affiliate, we earn on qualifying purchases.
From Knowledge Work to AI
The idea of a fourth sector emerged as economists and management thinkers tried to describe work that did not fit neatly into farming, manufacturing or conventional services. Peter Drucker, the Vienna-born management writer, used the term “knowledge worker” in 1959, in The Landmarks of Tomorrow. He argued that people increasingly earned a living through what they knew, rather than mainly through physical labor.
Machlup followed with an attempt to quantify knowledge-related activity. Since knowledge has no physical weight that can be counted like steel or grain, he measured its cost of production and distribution. His estimate captured spending across industries and occupations, but did not claim to measure the economic value of knowledge directly.
Porat’s later framework focused on what people did during their workday. A bookkeeper employed by a farm or a clerk in a steel mill could count as an information worker under that approach. The method helps explain why information work cuts across industries—and why AI’s potential reach cannot be assessed only by looking at technology companies.
Meyer places the historical figures alongside a current snapshot from the International Labour Organization. The article says managers, professionals and technicians account for 62% of jobs in Singapore, 52.5% in the United States and 47.9% in Germany, citing ILO data for 2025. Those occupational categories are not necessarily identical to Porat’s historical definition of information workers, so the figures should not be treated as a direct continuation of his series.
““knowledge worker””
— Peter Drucker, as described by Thorsten Meyer
What AI Exposure Shows So Far
The article does not identify the specific studies or methods behind its estimates that higher-income and clerical jobs are most exposed to generative AI. It also does not quantify how many jobs could be changed, displaced or created, or provide a timeline for any such effects. Exposure estimates are not forecasts of net job losses.
Meyer says overall U.S. job data shows no clear AI footprint so far, but provides no series or cutoff date for that assessment. The reported lag among younger workers in highly exposed jobs is also not accompanied by a named dataset or evidence that AI caused it. The scale and cause of that difference remain unclear.
Historical figures need similar care: Machlup measured spending, while Porat classified work by task. Their results illuminate the long rise of information activity but do not directly measure the present-day AI economy. The source also does not explain whether its current occupational shares use a definition comparable to either historical framework.
Evidence Needed on Job Effects
The next step is to track employment and workplace changes over time, using clearly stated definitions of AI exposure and job categories. That evidence would help distinguish tasks supported by AI from tasks replaced by it, and show whether effects differ by age, occupation or country.
Meyer’s article closes the first arc of Work After Work and points to a following episode about what economic measurements capture and what they leave out. For readers following this issue, the key development to watch is not exposure alone but whether subsequent data links AI adoption to measurable changes in hiring, pay, job duties or employment.
Key Questions
What does “fourth sector” mean?
It refers to knowledge and information work, alongside agriculture, industry and services. The category can include office and professional roles across many different industries.
What did Machlup’s 29% estimate measure?
Machlup estimated spending on producing and distributing knowledge in the United States in 1958: $136.4 billion, or almost 29% of adjusted GNP. It measured costs, not the direct value of knowledge.
Does AI exposure mean a job will be eliminated?
No. Exposure indicates that work may involve tasks generative AI could affect; it does not establish that a job will disappear or that employment will decline.
Has AI already changed overall U.S. employment?
Meyer’s article says overall U.S. job data has not shown a clear AI footprint so far. It does not provide the underlying data or a cutoff date, so the statement is a limited assessment, not proof that no effects exist.
Source: ThorstenMeyerAI.com
