🔍 Read the full analysis: The Fourth Sector's AI Opportunity 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 November 10, 2026 essay on ThorstenMeyerAI.com argues that information and knowledge work form a fourth economic sector—and that generative AI is aimed heavily at those office tasks. The historical figures are based on studies by Fritz Machlup and Marc Porat; the essay also says current evidence does not show a clear overall U.S. employment effect from AI.
ThorstenMeyerAI.com published an essay on November 10 arguing that generative AI is directed at work in the information and knowledge economy, a large part of employment that earlier economists measured as a distinct sector. The article connects historical estimates of knowledge work with current concerns about AI exposure, while stressing that exposure is not proof of job loss.
The essay centers on economist Fritz Machlup’s 1962 estimate that producing and distributing knowledge accounted for $136.4 billion, or “almost 29 per cent of adjusted GNP,” in the United States in 1958. Machlup measured spending on knowledge-related activities, rather than the value of knowledge itself. The article says education represented 44% of the total, media 28%, information services 13%, research and development 8%, and information machines 7%, citing a breakdown tabulated by historian of statistics Benoît Godin.
It also recounts Marc Porat’s four-sector classification of the U.S. workforce into agriculture, industry, services and information. In Porat’s 1977 report, information workers accounted for 5.8% of the workforce in 1860, 30.8% in 1950 and 46.4% in 1970. Porat classified people by what they did, not simply by their employer: a clerk in a steel mill, for example, counted as an information worker.
For current comparisons, the essay cites 2025 International Labour Organization figures showing managers, professionals and technicians held 62% of jobs in Singapore, 52.5% in the United States and 47.9% in Germany. It says higher-income and clerical jobs are among those most exposed to generative AI, but reports no clear AI effect in overall U.S. employment data so far. It also describes young workers in the most exposed jobs as falling behind peers; the source does not provide the underlying dataset or measures in the supplied text.
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
AI Reaches the Information Workforce
The article’s central point is that AI’s potential reach extends beyond factories and manual labor. Generative AI tools are designed to perform or assist with tasks common in office jobs, including writing, reading, calculating, checking and organizing information. Those activities have been part of economic production for decades, even when they do not produce a physical object.
The historical numbers help explain the scale of the issue: information work became a major share of the U.S. workforce during the twentieth century, while current employment figures cited by the essay show large proportions of jobs in professional, managerial and technical categories. If AI changes how those tasks are done, the effects could reach many workers and employers. But the article does not establish that exposure will lead to layoffs, lower pay or reduced hiring. Potential task automation and measured employment outcomes are different questions, and the source says a broad U.S. employment effect is not yet clear.
Products Worth Considering
Anker SOLIX C300 DC Portable Power Station with 60W Solar Panel, 288Wh, 300W
Samsung Galaxy Tab A11+ (6GB RAM, 128GB Storage) – Gray
Amazon Kindle Paperwhite (2024, 16 GB) – 7-inch Glare-Free Display, Black
As an affiliate, we earn on qualifying purchases.
From Knowledge Work to AI
Management writer Peter Drucker used the term “knowledge worker” in 1959, according to the source article, to describe people earning a living through what they know rather than primarily through manual labor. Drucker’s argument, as presented there, was that managing knowledge workers would require approaches different from factory production methods such as the assembly line.
Machlup tried to quantify knowledge-related activity by counting its costs across education, research, media, information machinery and services. Porat later tracked information workers by occupation and expanded the picture into a four-sector framework. The essay calls this broad category the quaternary, or fourth, sector. The measures are not interchangeable: Machlup estimated spending, while Porat classified workers and also examined income and value added.
The piece is the final installment in the first arc of Thorsten Meyer’s “Work After Work” series. Its historical discussion provides context for the AI question, but the old estimates do not by themselves measure today’s AI adoption, productivity or employment effects.
Exposure Is Not an Employment Result
The supplied article does not identify the studies or methods behind its statements about generative AI exposure, the absence of a clear overall U.S. employment footprint, or younger workers in highly exposed jobs falling behind their peers. It gives no time window, comparison group or definition of “falling behind” for that last claim. Those claims should therefore be read as the article’s account, not as fully documented findings in the material provided.
It is also unclear which specific jobs or tasks will be automated, how widely employers will adopt AI, and whether adoption will replace work, change job duties or increase demand for some roles. Historical estimates of knowledge work establish the sector’s scale, but they cannot predict the size or timing of AI-related employment changes. No quantified forecast of job losses or gains is provided.
Evidence to Watch in Employment Data
The next test is whether employment, hiring, wages and job duties in highly exposed occupations change in ways that can be tied to AI use. The source article says the overall U.S. job data does not yet show a clear AI footprint, but does not specify which datasets or future release dates readers should track.
For now, the article frames the fourth sector as the part of the economy where AI’s effects may be especially relevant, rather than claiming that those effects are settled. More evidence on workplace adoption and worker outcomes will be needed to distinguish task exposure from actual changes in employment.
Key Questions
What is the “fourth sector”?
It is the information or knowledge sector, covering work that produces, processes or distributes information. The article describes it alongside agriculture, industry and services.
What did Machlup’s 29% figure measure?
It was Machlup’s estimate that knowledge-related production and distribution represented almost 29% of adjusted U.S. GNP in 1958. The figure measured related spending, not the value of knowledge itself.
Does AI exposure mean a job will disappear?
No. The source explicitly distinguishes exposure to generative AI from job loss. It says there is no clear overall AI footprint in U.S. employment data so far.
How large was information work in Porat’s estimate?
Porat’s 1977 report classified 46.4% of the U.S. workforce as information workers in 1970, up from 5.8% in 1860, according to the figures reproduced in the article.
What evidence is still needed?
More specific data is needed on AI adoption, changes to job tasks, hiring, wages and employment, including details behind the article’s claim about younger workers in highly exposed jobs.
Source: ThorstenMeyerAI.com
