🔍 Read the full analysis: Knowledge Work Insights On AI 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
ThorstenMeyerAI.com’s Nov. 10 report examines how knowledge and information work became a major part of the U.S. economy, drawing on estimates by Fritz Machlup and Marc Porat. It says generative AI is aimed at many office and professional tasks, but exposure does not establish that jobs will disappear; the report cites no clear AI effect in overall U.S. employment data so far.
ThorstenMeyerAI.com published a report Nov. 10 on the rise of knowledge work and its relevance to generative AI, arguing that the technology is directed at many tasks carried out in offices, including writing, calculating, checking and planning. The report draws on historical estimates that information-related activity had become a major part of the U.S. economy decades ago, while stressing that AI exposure is not the same as job loss.
The report, part of the eighth episode of the series Work After Work, centers on two efforts to measure the knowledge economy. Economist Fritz Machlup estimated that U.S. spending on producing and distributing knowledge reached $136.4 billion in 1958, or “almost 29 per cent of adjusted GNP.” The report notes that Machlup measured spending costs, not the value of knowledge itself.
Machlup grouped knowledge-related activity into education, research and development, media, information machines and information services. A breakdown cited in the report, based on historian Benoît Godin’s tabulation, assigns 44% of the total to education, 28% to media, 13% to information services, 8% to research and development, and 7% to information machines. The report says knowledge industries grew 8.8% annually from 1947 to 1958, according to Machlup’s estimate.
The second historical measure comes from Marc Porat’s 1977 study for the U.S. Department of Commerce. Porat classified workers by the tasks they performed rather than the industry where they worked. His figures put information workers at 5.8% of the U.S. workforce in 1860 and 46.4% in 1970. The report also cites 2025 International Labour Organization figures showing managers, professionals and technicians accounted for 52.5% of U.S. jobs, 47.9% in Germany and 62% in Singapore, amid broader changes in work as AI takes hold. These categories are not identical to Porat’s definition of information work.
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 Office Work Is in Focus
The report’s central point is that AI is aimed at tasks that already occupy a large share of work, not only at physical labor. Generative systems can assist with or automate parts of reading, drafting, summarizing, calculation and information checking. That makes their potential effects relevant to clerical and professional workers, as well as to employers deciding how to organize office work.
But measuring exposure is not a forecast of how many jobs will disappear. A job may include tasks that AI can assist with alongside work that still requires human judgment, interaction or accountability. The report says higher-income and clerical roles appear near the front of the line for generative AI exposure, while distinguishing that assessment from evidence of actual job losses.
It also cites research indicating that overall U.S. employment data has not yet shown a clear AI footprint. The report flags a possible exception: young workers in the most exposed jobs are falling behind their peers. It does not provide enough detail here to establish the size, cause or duration of that gap. The distinction matters for readers: the broad labor-market impact remains unsettled, even as particular workers may be experiencing changes.
Popular right now
Roborock RockNeo Q110H Wire-Free Robot Lawn Mower
Vitamix Touch & Go Advance 32-Ounce Bar Blender, White
Breville Barista Express Espresso Machine with Built-In Grinder, Black Sesame
As an affiliate, we earn on qualifying purchases.
From Knowledge Workers to AI
Management writer Peter Drucker described the rise of “knowledge work” in his 1959 book The Landmarks of Tomorrow. The report credits Drucker with popularizing the term “knowledge worker” for people who earn a living using what they have learned. His argument, as presented in the episode, was that organizations could not rely only on factory-era methods such as assembly lines when more work depended on knowledge.
Machlup attempted to quantify this activity in The Production and Distribution of Knowledge in the United States, published in 1962. Porat later expanded the approach by classifying workers according to what they did during the workday. The report uses their work to frame information as a distinct economic sector alongside agriculture, industry and services, sometimes called the quaternary sector.
The historical estimates use different methods and periods, so they should not be treated as a single continuous series. Machlup counted spending across knowledge industries; Porat classified occupations and economic activity. Their work establishes the long history of information-intensive labor, but it does not by itself measure the current scale or effects of generative AI.
“almost 29 per cent of adjusted GNP”
— Fritz Machlup, as quoted in the report
What AI Exposure Shows So Far
The report does not give a specific estimate of how many jobs generative AI will eliminate, change or create, nor does it identify a timeframe for those outcomes. Exposure estimates describe possible task overlap; they do not confirm that employers have adopted AI at scale or that workers have been displaced because of it.
The cited claim that young workers in highly exposed jobs are falling behind their peers is not accompanied in the supplied material by a study title, sample, measurement period or numerical result. Its cause is also not established. The report’s statement that overall U.S. employment data shows no clear AI footprint is similarly presented without a named dataset or analysis in the episode text.
Historical shares also depend on definitions. The Machlup and Porat figures measure different things, and the modern ILO categories of managers, professionals and technicians are not a direct equivalent to Porat’s information-worker group. Readers should treat the figures as indicators of a broad shift, not as directly comparable measures.
The Series Turns to Measurement
The episode identifies the next installment as a discussion of what economic measures capture and what they leave out. That question follows directly from Machlup’s method: his 29% figure measured costs associated with producing and distributing knowledge, not the full value or quality of the knowledge produced.
For the AI labor question, the next useful evidence will be research that tracks adoption, changes in specific tasks, hiring and employment over time, with clear comparisons between more- and less-exposed roles. The report does not announce a date for that evidence or provide a forecast. For now, it presents the historical growth of information work as context for a developing debate, not proof of a particular employment outcome.
Key Questions
What did the report say happened?
ThorstenMeyerAI.com published an episode on Nov. 10 tracing the growth of knowledge and information work and explaining why generative AI may affect office tasks.
What does the 29% figure measure?
It is Fritz Machlup’s estimate that U.S. spending on producing and distributing knowledge in 1958 amounted to $136.4 billion, or almost 29% of adjusted GNP. It measures spending, not the total value of knowledge.
Does AI exposure mean jobs will be lost?
No. Exposure means tasks may overlap with AI capabilities; it does not establish that a job will disappear. The report says overall U.S. employment data has not shown a clear AI effect so far.
How did Porat measure information work?
Marc Porat classified workers by what they did at work, rather than simply by their employer’s industry. His 1977 study estimated that information workers made up 46.4% of the U.S. workforce in 1970.
What remains unknown about AI’s effect on workers?
The report does not specify how many jobs might change or disappear, over what period, or how much of the change would be caused by AI. It also gives limited detail for its claim about young workers in highly exposed jobs.
Source: ThorstenMeyerAI.com
