TL;DR
A November 10, 2026, essay revisits estimates that knowledge-related activity accounted for nearly 29% of adjusted U.S. GNP in 1958 and that information workers had become the largest U.S. workforce group by the mid-1950s. It connects that history to generative AI’s exposure of office work, while noting that exposure does not establish job loss and that broad U.S. employment data has not shown a clear AI effect so far.
The essay anchors its account in economist Fritz Machlup’s 1962 study of knowledge production in the United States. For 1958, Machlup estimated that spending on producing and distributing knowledge totaled $136.4 billion, or “almost 29 per cent of adjusted GNP.” The figure measured spending, not the value or quality of the knowledge produced.
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
Machlup grouped knowledge industries into education, research and development, communications media, information machines and information services. As tabulated by historian of statistics Benoît Godin, education represented $60.2 billion, or 44% of the total; media accounted for $38.4 billion, or 28%. The essay also reports Machlup’s estimate that workers in knowledge occupations rose from about 11% in 1900 to 32% in 1959.
The article connects those measures to later work by Marc Porat, who classified workers by their daily tasks. It says information workers grew from less than 6% of the U.S. workforce in 1860 to 46% in 1970, becoming the largest worker group by the mid-1950s. For a current comparison, it cites 2025 International Labour Organization figures showing managers, professionals and technicians made up 62% of jobs in Singapore, 52.5% in the United States and 47.9% in Germany.
Knowledge Work Meets AI in the Fourth Sector
A November 2026 essay connects the long rise of information work to generative AI’s reach into office tasks. Exposure signals potential change, not confirmed job loss.
How knowledge work grew
Different historical measures show the expansion of work centered on knowledge and information.
AI reaches across office tasks
Writing, planning, checking and communication are common in jobs that handle information.
Clerical work
The essay says clerical jobs may be near the front of the line for generative AI exposure.
Higher-income roles
Some estimates also place higher-income occupations among those with greater exposure.
Tasks may change
AI may assist with tasks, automate some of them or have little effect on a particular job.
Exposure is not job loss
The essay distinguishes possible task overlap from demonstrated employment effects.
- Some estimates flag clerical and higher-income jobs as more exposed.
- Broad U.S. employment data has not shown a clear AI footprint so far.
- The size, timing and distribution of future employment changes.
- Whether AI caused the reported lag among younger workers in exposed jobs.
Evidence to watch
Useful comparisons need clear time spans, job groups and measures.
- Track hiring and job changes in occupations with high estimated exposure.
- Measure workplace task adoption over time.
- Separate AI-related changes from other economic forces.
AI’s Reach Across Office Work
The historical figures help explain why AI adoption has drawn attention beyond technology companies: many jobs involve handling information, including writing, planning, checking and communication. The essay says estimates place clerical and higher-income occupations near the front of the line for generative AI exposure, suggesting that the technology could affect tasks common in large parts of the labor market.
That exposure does not by itself show that workers will be replaced. A tool may automate some tasks, change how they are done or have little effect on a particular job. The article says 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. It gives no specific measure, time period or causal evidence for that comparison, so it should be read as a reported signal rather than proof that AI caused the gap.
The distinction matters for employers and workers deciding how to adapt. Counts of knowledge occupations describe the scale of the workforce involved; they cannot, on their own, predict how quickly AI will change work, which tasks will be affected or whether new work will offset losses.
Popular right now
De'Longhi Classic Signature Espresso Machine with Adjustable Milk Frother
De'Longhi La Specialista Touch Semi-Automatic Espresso Machine with Grinder
TCL 65-Inch 4K UHD Smart LED TV (65S425)
As an affiliate, we earn on qualifying purchases.
From Knowledge Work to Information Jobs
Management writer Peter Drucker described the rise of knowledge work in his 1959 book The Landmarks of Tomorrow. As presented in Meyer’s essay, Drucker’s central point was that organizations built around factory production might not be suited to workers whose output depends chiefly on learned expertise and judgment.
Machlup then tried to quantify this changing economy by tracking industries and occupations associated with creating and sharing knowledge. His 29% estimate is specific to U.S. spending in 1958 and to his definition of knowledge production; it is not a direct measure of today’s economy. Porat’s later count used a different approach, sorting people by the information tasks they performed.
The measures are useful historical markers, but they are not interchangeable. The essay’s modern country comparisons use occupational categories for managers, professionals and technicians, while its discussion of AI exposure concerns the tasks generative AI may affect.
“almost 29 per cent of adjusted GNP”
— Fritz Machlup, as quoted in his 1962 study
Employment Effects Still Unclear
The essay does not identify the specific studies behind its estimates that higher-income and clerical jobs face greater generative AI exposure, so the methods and definitions behind those rankings are not detailed here. Exposure can refer to tasks that AI may assist with or automate; it does not establish that an occupation or job will disappear.
It is also unclear what time span or data supports the reported lag among younger workers in the most exposed jobs, and whether AI is responsible. The essay says broad U.S. job data has no clear AI footprint so far, but does not specify the dataset or the threshold used to identify one. The scale, pace and distribution of any future employment effects remain unsettled.
Evidence to Watch in Hiring
The next evidence will come from employment and workplace data that track hiring, job changes and task adoption over time, particularly in occupations with high estimated exposure. Comparisons will need to distinguish changes associated with AI from other economic forces and specify the dates, job groups and measures being used.
Meyer frames the essay as the final episode of the first arc of Work After Work and points to a following installment about what economic statistics measure and leave out. For the AI question, the key open issue is whether exposure translates into measurable changes in work and employment.
Key Questions
What is the essay’s main news development?
A November 10, 2026, essay connects the long rise of knowledge work to generative AI’s focus on office tasks and summarizes what is currently known about job exposure.
What did Machlup estimate for 1958?
He estimated that knowledge production and distribution accounted for $136.4 billion, or almost 29% of adjusted U.S. GNP. This was a spending estimate, not a measure of knowledge’s value.
Does AI exposure mean jobs will be lost?
No. Exposure describes potential overlap between AI capabilities and job tasks. It does not establish that workers will be displaced, and the essay says broad U.S. employment data has not shown a clear AI effect so far.
Which jobs does the essay say may be more exposed?
It says estimates put clerical and higher-income jobs toward the front of the line for generative AI exposure. The essay does not provide the underlying studies’ methods or detailed job-level findings.
What remains unknown about AI’s labor-market effects?
The size, timing and distribution of any employment changes remain unclear. The essay also does not provide enough detail to assess its report that younger workers in the most exposed jobs are falling behind their peers or determine whether AI is a cause.
Source: Thorsten Meyer AI
