🔍 Read the full analysis: The Fourth Sector: AI's New Frontier on ThorstenMeyerAI.com
Get smart everyday buys 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 from ThorstenMeyerAI.com traces the rise of knowledge and information work as a fourth economic sector, drawing on studies by Fritz Machlup and Marc Porat. It connects that history to generative AI’s exposure of office and professional work, while stressing that exposure is not the same as job loss and that broad US employment data has not yet shown a clear AI effect.
ThorstenMeyerAI.com published an episode on November 10 examining how knowledge and information work became a major part of the economy—and why those jobs are now a focus of generative AI. The report draws on economist Fritz Machlup’s estimate that knowledge-related industries accounted for almost 29% of adjusted U.S. GNP in 1958, and Marc Porat’s later finding that information workers made up 46.4% of the U.S. workforce in 1970.
The episode, the last installment in the publication’s first arc of its “Work After Work” series, describes a fourth, information-based sector alongside agriculture, industry and services. Its central point is that a large share of work involves producing, processing or distributing information, even when those jobs are located in factories, farms or other sectors.
Machlup’s 1962 book, The Production and Distribution of Knowledge in the United States, estimated that U.S. spending on knowledge production and distribution reached $136.4 billion in 1958, or “almost 29 per cent of adjusted GNP.” The episode says this was an expenditure measure: Machlup counted the cost of producing knowledge, not its market value. Education represented $60.2 billion of the total, according to a breakdown attributed in the source to historian of statistics Benoît Godin.
Porat’s 1977 study for the U.S. Department of Commerce classified workers by their tasks rather than their employer’s industry. It estimated that information workers accounted for 5.8% of the U.S. workforce in 1860 and 46.4% in 1970. The episode says the group had become the largest employment category by the mid-1950s. It also cites 2025 International Labour Organization figures showing managers, professionals and technicians made up 52.5% of jobs in the United States, 47.9% in Germany and 62% in Singapore.
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 Office Work
The historical account matters because generative AI is aimed at tasks common in knowledge jobs: writing, reading, calculating, checking and communicating. The source says estimates place higher-income and clerical occupations among those most exposed to generative AI. That could affect a substantial part of the workforce, not only technology firms or workers in jobs commonly associated with automation.
Exposure, however, is not a measure of how many jobs will disappear. AI may change tasks within roles, alter how much work is needed, or create new work; the episode does not establish which outcome will prevail. It reports that overall U.S. employment data has not yet shown a clear AI footprint, while younger workers in the most exposed jobs are falling behind their peers. Those observations signal possible uneven effects, but they do not by themselves show that AI caused the gap.
The distinction is relevant for workers and employers deciding how to respond. A job title alone may not reveal exposure: Porat’s task-based approach counted, for example, a clerk in a steel mill as an information worker. The same lens can help explain why AI’s reach may cross industry boundaries, while still leaving the scale and timing of employment effects uncertain.
Products Worth Considering
Dell UltraSharp U3225QE 31.5-Inch 4K UHD 120Hz IPS Black Thunderbolt Hub Monitor
Samsung 32 Odyssey QD-OLED G8 Gaming Monitor – 4K, 240Hz
Dell 32 Plus 4K Monitor S3225QS, 31.5-Inch VA, 120Hz, FreeSync Premium
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 episode, to describe people earning a living through what they know rather than primarily through manual labor. Drucker argued that as work shifted toward creating value with knowledge, factory-era management methods would not necessarily fit. The source presents that as an early account of a changing workforce, not a forecast of today’s specific AI systems.
Machlup then tried to quantify the knowledge economy by grouping education, research and development, communications media, information machines and information services. Porat later extended the approach by classifying occupations according to what workers did. His figures show the information category rising over time as agriculture’s share fell and industry’s share eventually declined. The episode uses these studies to frame the quaternary, or information, sector as a longstanding economic development rather than something created by AI.
The measures are not interchangeable. Machlup estimated spending on selected industries, while Porat classified workers and estimated employment and income shares. The modern figures cited for Singapore, the United States and Germany count managers, professionals and technicians, a different occupational grouping. Together they offer historical and present-day indicators of knowledge work, not one continuous measure of the same sector.
““almost 29 per cent of adjusted GNP””
— Fritz Machlup, as quoted in the episode’s account of his 1958 estimate
What AI Exposure Does Not Show
The episode does not identify the underlying studies behind its estimates of which jobs are most exposed to generative AI, or provide a single exposure figure for the workforce. It also does not specify the period or method behind the observation that younger workers in the most exposed jobs are falling behind their peers. The relationship between that pattern and AI is therefore unclear from the material provided.
There is no confirmed estimate here of how many jobs AI has eliminated, how many it has created, or how quickly employment could change. The historical figures also rely on definitions that differ by study and period. They show the growth of information-related work under those definitions; they do not directly predict the future effects of generative AI.
What the Evidence Must Track
The episode closes the publication’s first arc and points to a following installment about what economic numbers measure and what they leave out. That topic bears directly on interpreting AI’s effects: spending on technology, job counts and changes in work tasks answer different questions.
For a clearer picture of AI’s labor-market impact, future evidence will need to track employment and wages over time, distinguish job losses from changes in duties, and compare more- and less-exposed workers while accounting for other causes. The source identifies no new government release, research deadline or specific policy decision. For now, the status it describes is mixed: substantial exposure is anticipated in office work, but a broad, measurable effect on U.S. employment has not been established.
Key Questions
What does the episode mean by the fourth sector?
It means information or knowledge work, treated as a sector alongside agriculture, industry and services. The episode describes workers who produce, process or distribute information, including people employed outside industries normally thought of as office-based.
What does Machlup’s 29% figure measure?
It is Machlup’s estimate of spending on knowledge production and distribution in the United States in 1958: $136.4 billion, or almost 29% of adjusted GNP. It measures cost, not the total value of knowledge.
Does AI exposure mean a job will be eliminated?
No. Exposure describes a job’s potential contact with AI-related tasks; it does not establish that the job will disappear. The source says overall U.S. job data has not yet shown a clear AI footprint.
What evidence of AI’s employment effects does the source report?
The episode says broad U.S. employment data shows no clear AI effect so far, while younger workers in the most exposed jobs are falling behind their peers. It does not establish that AI caused that difference or provide a precise estimate of job losses.
Source: ThorstenMeyerAI.com
